Scholia Reviews ns 8 (1999) 13.

Sheila Murnaghan & Sandra R. Joshel (edd.), Women and Slaves in Greco-Roman Culture: Differential Equations. Routledge: London and New York, 1998. Pp. xii + 288. ISBN 0-415-16229-7. UK£45.00.

Marc Kleijwegt
Department of Classics, University of South Africa

Since the 1970s the study of women in the ancient world has been one of the most excitingly innovative fields in Classics. The shift from a traditional study of legal status and related problems to that of personal experiences required more sophisticated ways of enquiry. More than any other modern focus on the ancient world, women's studies has raised the question of the nature of our evidence, underlining the male authorial bias of the literary texts and the unavailability of personal documents. Consequently, to get meaningful results, women's studies, or at least a major branch of it, has been characterized by a strong inclination to use interdisciplinary and comparative methods, accompanied by a less pronounced but still significant focus on non-literary evidence, such as inscriptions and papyri. The study of ancient slavery is beset with similar problems and has seen a similar rise in the value of the comparative approach, although it is as yet not as prominent as in women's studies. To combine the two fields is therefore a most welcome step and this volume of essays should be regarded as a good first attempt to sensitise scholars to the interconnectedness of gender and slavery. Such a connection has been largely ignored up till now, for as one of the contributors, Nancy Rabinowitz, remarks, there has been a strong tendency to study women without paying attention to class and to study slavery without considering gender (p. 57). The emphasis in the essays is on cultural representations of women and slaves, or in other words, on how women and slaves were construed as parts of a symbolic discourse maintained by a small group of elite men. As the editors put it in their introductory chapter, gender and slavery are not independent phenomena, 'but intersecting variables in a process that we have labeled "differential equations" whereby women and slaves are assimilated only to be distinguished, compared but never quite identified' (p. 3).

The lengthy introductory chapter by the editors is followed by thirteen essays, six of which deal with the Greek world and the remainder discuss Roman topics. Since I will not be able to discuss all of them in the same detail, I provide a run-down of the contents for the benefit of the reader: Sandra R. Joshel and Sheila Murnaghan, 'Introduction: Differential Equations' (pp. 1-22); William G. Thalmann, 'Female Slaves in the Odyssey' (pp. 22-35); Denise McCoskey, '"I, Whom She Detested so Bitterly": Slavery and the Violent Division of Women in Aeschylus' Oresteia' (pp. 35-56); Nancy Rabinowitz, 'Slaves with Slaves: Women and Class in Euripidean Tragedy' (pp. 56-69); Nancy Demand, 'Women and Slaves as Hippocratic Patients' (pp. 69-85); Richard P. Saller, 'Symbols of Gender and Status in the Roman Household' (pp. 85-92); Annalisa Rei, 'Villains, Wives and Slaves in the Comedies of Plautus' (pp. 92-109); Patricia Clark, 'Women, Slaves and the Hierarchies of Domestic Violence: The Family of St Augustine' (pp. 109-30); Joy Connoly, 'Mastering Corruption: Constructions of Identity in Roman Oratory' (pp. 130-52); Holt Parker, 'Loyal Slaves and Loyal Wives: The Crisis of the Outsider- within and Roman Exemplum Literature' (pp. 152-74); Kathleen McCarthy, 'Servitium Amoris: Amor Servitii' (pp. 174-93); Ian Morris, 'Remaining Invisible: The Archaeology of the Excluded in Classical Athens' (pp. 193-221); Steven Johnstone, 'Cracking the Code of Silence: Athenian Legal Oratory and the Histories of Slaves and Women' (pp. 221-36); Shane Butler, 'Notes on a Membrum Disiectum' (pp. 236-56).

In this review the main focus will be on a series of issues in connection with the usefulness of the volume's approach, especially its emphasis on literary texts and symbolic discourse, and on the value of studying cultural representations as opposed to a straightforward historical study. The volume contains many good points, but it is in a sense the victim of a lack of a body of innovative research in the field of slavery. By including slavery in their discussion some of the contributors are clearly entering less familiar territory, and their arguments are sometimes not fully supported by what historical studies have taught us. Yet an innovative collection such as this one deserves to be judged on the benefits it may offer to both scholars and students. These are plentiful and easy to pick out, so if the following remarks and observations tip the balance towards the negative, this is done for the reason that the setting of the methodological framework requires a fuller discussion of its principles and possibilities than is offered in the volume itself. I shall illustrate these points in a detailed discussion of a few of the chapters. My selection does not imply that these are necessarily the weakest chapters in the volume. They were chosen for the reason that they provide a good illustration of the shortcomings which I have signalled. They are also, in my view, the chapters which, because of their ingenious ways of trying to find new perspectives, are most likely to stimulate further research.

The central axis around which the essays are organized is the way in which individual texts create representations of women and slaves (in fact, Richard Saller and Ian Morris are the only contributors who do not frame their discussion around a central text). This is done either by presenting them in separate discussions (Demand; Parker) or by focusing on their interconnectedness in one specific text (the overwhelming majority). What is quite evident is the explicit distinction most contributors draw between 'historical' and 'cultural' women and slaves. More than once a historical perspective is mooted to be instantaneously differentiated from the approach selected by the author. Thus, the study of cultural representations is seen in terms of a project which is mainly grounded in a sensitised reading of literary texts, while the study of historical features must be performed on the basis of different premises and with an entirely different spectrum of source-material. This extreme position is unnecessarily confusing, for it has not stopped contributors from making a connection between individual passages and developments in society. How such a connection can be identified successfully, and whether the author's conclusions are indeed correct, is more problematic than is sometimes realised. As the editors argue, the Greeks and Romans made frequent use of the polarities slave/free and male/female to understand themselves and to organize their society (p. 3). This argument presupposes that the same rules which underlie the creative imagination of the two cultures can be applied to the social framework of these societies. In fact, they may have been completely different. As is clear from Richard Saller's discussion, in social terms the Roman household was based in and functioned on the distinction between slave and free. If the married upper-class woman is sometimes seen acting in a way that blurs the line between herself and her female slaves, this must be viewed as the outcome of a symbolic action which is only valid for the duration of a specific festival. The circumstances in which this 'égalité volontaire' is performed make it clear that in a normal context there was no lack of clarity about the respective statuses of free matronae and female slaves. Saller's position stands in stark contrast to that of the other contributors who focus their attention on how in literature women and slaves are frequently compared in respect of their position vis-à-vis the paterfamilias. The discussion operates on different levels, taking their starting point from opposite ends of the spectrum on which women and slaves were measured.

The split between a historical and a cultural study of women and slaves is made most explicitly in Denise McCoskey's chapter on the confrontation between a free woman (with male aspirations) and a female slave (but formerly of high birth) in Aeschylus' Agamemnon. Her approach deserves detailed attention all the more because her choice in drawing on comparative material is an intrinsically fruitful one. McCoskey offers the following justification for preferring fictional over historical evidence: 'such a complex articulation of the relationship between slave and free woman is difficult to find in historical evidence from ancient Greece (a condition hampered by the lack of extant writing by women and slaves)' (p. 36). The second part of her argument is put in parentheses and thus must be understood as less significant than the main point she wants to make (which is a valid one, for constructed realities on occasion point up a society's possibilities and impossibilities in significant ways). Yet the matter deserves some attention, for it appears in most of the chapters. The observation presupposes that such writings, if they had been extant, would have presented us with a radically different point of view from what is available through material written by elite men. I believe that the implied optimism that there once existed a corpus of ancient literature which contained a potentially more authentic expression of the female or the slave experience is largely anachronistic. No significant writings by members of these groups have come down to us, for the simple reason that there was not very much to begin with.[[1]] I therefore find extremely plausible Ian Morris's argument that male Athenians also dominated the material world to such an extent that it is unsurprising that archaeologists have found little trace of women and slaves. Not being an archaeologist myself, I do not possess the knowledge to test Morris's argument, but what I do know is that the study of epigraphy, even though the material appears to be more authentic, does not provide a less problematic route to the experiences of women and slaves.[[2]]

The fact that in Aeschylus' tragedy the relationship between Clytemnestra and Cassandra is articulated in such a complex way makes one wonder at what level of discourse it should be located. Like any piece of good literature the Agamemnon relies not on static concepts, but on a wide variety of ways and means to surprise the spectator, including the collapsing of polarities. The original, stark contrast between the two women is gradually elided. This is done, for instance, by making Clytemnestra address Agamemnon in a foreign style (thereby adopting the aura of foreignness which she herself tries to stamp on Cassandra) and by allowing Cassandra to speak perfect Greek in spite of her presumed barbaric origins. It is significant that Clytemnestra displays her foreign characteristic before she has met Cassandra, suggesting perhaps that her past criminal behaviour and her premeditated plan to murder her husband are responsible for this change in ethnicity. During her first encounter with Cassandra Clytemnestra is therefore displacing her acquired foreign status onto another woman whom she is at liberty to brand as a foreigner by birth.

By drawing on the writings of black female slaves from the American South McCoskey adds a fascinating comparative (and ultimately political) dimension to her discussion. These writings vividly express the lack of solidarity between white mistresses and black female slaves. It is striking to note that, although in their oppressed situation they shared many similarities, this is only acknowledged by one of the parties. The white mistresses refused to be drawn into what they may have considered as a disturbing set of parallels and they responded by emphasizing the alterity of the slaves. Such a response, set in a racially divided society, would seem to offer an unpromising parallel for an understanding of the cultural polarisation in the Agamemnon. Yet McCoskey makes an interesting case to identify a similar function for the emphasis on ethnicity in Aeschylus. She explicates along these lines Clytemnestra's stress on Cassandra's inability to reply to her as a lack of civilisation (1059-61 must be taken to mean something like: 'if you cannot understand what I'm saying, just wave your little barbarian hand'). It is clear, however, that the polarity of slave and free is blended with traditional ploys of tragedy (Cassandra is allegedly unable to understand what is going on, although she knows perfectly well what is going to happen), which makes it difficult to establish the value of the confrontation as a reflection of fifth-century domestic predicaments. The inclusion of comparative material is a useful means to provide a plausible reconstruction of experiences for which the ancient evidence proves insufficient. Especially in the study of ancient slavery the comparative method has been used with reasonable success. However, McCoskey's use of it here raises some important questions. She briefly defends her choice of such material (p. 37), sketching in broad lines the similarities between ancient Greek slavery and slavery in the American South. Yet these parallels merely concern the system of slavery adopted in both societies, a convergence which does not imply that the human responses were always the same. One important difference lies in the position of American slavery in its contemporary global context and the response which it evoked. Slaves in the American South were writing down their experiences in a world which regarded slavery as a social evil.[[3]] The ancient world, however, never abolished slavery. Although it was a point of debate in philosophical discourse, slavery was never actively opposed or even condemned by more than a few isolated individuals. In such a society, I would argue, there was no established place for the recording of slave experiences, for who would be interested in the slave's point of view? In other words, the existence of an ideological framework which condemned slavery rather than supported it made possible the conceptualisation of sentiments which may have been unrecognisable in those terms to masters and slaves in the ancient world. Even more uncertain is the point as to whether Aeschylus can be depended upon to have identified with the feelings of a slave, and not with those of a high-born aristocrat whose fate it was to have become a prisoner of war, a slave and a concubine.

Do the parallels which McCoskey signals adequately reflect the plot and language of the Agamemnon? The most striking similarities are those between the white mistresses' emphasis on the alterity of their black slaves and Clytemnestra's attempts to brand Cassandra as a foreigner, but they are less pronounced in the area of shared suffering. McCoskey attempts to show that the two women, in addition to being images of oppression and resistance, are linked in more positive ways. The points of convergence, their mutual use of a cryptic style of communication and Cassandra's singular ability to comprehend Clytemnestra's speech (p. 43), are significant elements in the unfolding of the tragedy, but they do not support the idea of shared oppression. Cassandra's part is not that of a woman reaching out in vain to another woman, an effort which is then doomed by the stern hostility of Clytemnestra. As has been argued recently, her part may be that of a male focal point constructed in opposition to Clytemnestra's criminal behaviour and as such it illustrates a subjective viewpoint which seeks the approval of a male audience.[[4]] McCoskey ends her essay with a series of observations on how feminist readings of the Agamemnon, by ignoring Clytemnestra's victimisation of Cassandra, have served a white European middle-class brand of feminism (p. 52). The point itself is a useful reminder that all our readings of ancient texts are informed by specific versions of ideology. Yet McCoskey's view of the Agamemnon is just as much dictated by political considerations and may equally not do full justice to the play's intentions. From a methodological point of view I have great difficulty in accepting the validity of a comparison between a male-authored tragedy and autobiographical writings from a much later period in history. And I have my doubts especially where the comparison has the effect of making the confrontation between Clytemnestra and Cassandra into an authentic encounter. McCoskey's essay is a passionate and well- written effort to introduce a much-needed corrective to women's studies, but the text which she has selected as the vehicle for her arguments may be part of male fictions about women rather than an independent testimony as to how the latter experienced slavery.

McCoskey's essay is followed by Nancy Rabinowitz's discussion of slaves by convention, natural slaves and the role of gender in Euripides' Trojan War plays. The pairing of these two essays is an interesting one, not in the least by the different emphases which are given to the dynamics of the plays. This has as much to do with the imprints which the playwrights have stamped onto their subject- material as with the ideological viewpoints of the two scholars who study them. While McCoskey is inclined to view Cassandra's role as carrying the possibility of an authentic response, Rabinowitz's much broader-based approach places a stronger emphasis on how the plays respond to the desires and anxieties of the author and his audience (p. 66; emphasis as in the original). She argues that the prominence of female slaves in Euripides serves to allay the male spectators' anxieties about their own slaves -- that they might be attacked by them -- and about the fragility of their own freedom -- that Greek men might be enslaved themselves. The latter fear, Rabinowitz argues, was not an imaginary one for enslavement was often the result of inter-city warfare in the fifth century. In her discussion of these themes Rabinowitz takes up a similar approach as McCoskey by studying the confrontation between Hermione and Andromache, where the poet reduces the antagonism between slave and free to sexual competition for men, an element which also features in the Agamemnon. Yet, in contrast to McCoskey, Rabinowitz also pays close attention to the role of (in Aristotle's terms) natural slaves. This category is represented, for instance, by Andromache's typically nameless servant, who is portrayed as continuing to serve her mistress loyally. Rabinowitz concludes that gender similarity is stressed to short-circuit any potential threat of slave resistance (p. 64). Rabinowitz's discussion adds an interesting perspective to the debate on Euripides' social sensibility, which is traditionally considered to have been more pronounced than that of Aeschylus and Sophocles. She shows the natural limitations to his social awareness, for his plays, although perhaps less blatantly than Aeschylus, still bear the male subjective trademark.

The connection between the text's alleged purpose and the audience's possible response (the slave-owners' feel-good factor) as envisaged by Rabinowitz is, however, not without problems. Anxieties about class conflict and fears of enslavement are difficult to measure and equally problematic is the assumption that both were sufficiently prominent in Euripides' time to be in need of allaying, irrespective of our willingness to be persuaded by the idea that tragedy was an effective instrument in that process.[[5]] It is a priori natural to expect that in Athens, as in Rome, slave-owners had an instinctive fear that their slaves would hate them and take any opportunity to revolt. It is rather unfortunate that Rabinowitz fails to provide any evidence in support of her theory, although such evidence does exist,[[6]] but even then we are still a long way from revealing the intentional or unconscious aims of Euripides' plays. Such doubts may especially be expressed for Rabinowitz's second set of anxieties. That male spectators were infused with fear about their own possible enslavement is hard to prove or disprove, and to present it as a fundamental grounding of a Euripidean tragedy requires a serious historical exegesis of the problem. In his Slavery in Ancient Greece, Rabinowitz's main source of information on slavery in Athens, Yvon Garlan has calculated that enslavement followed a victory in inter-city warfare in one-quarter of the cases reported by ancient historians.[[7]] Yet, the ideological matrix in which such acts were performed and eventually justified is a complex one. In the first place writers of the fifth and fourth centuries frequently expressed their horrors about Greeks being kept as slaves by other Greeks. Secondly, the worst effects of enslavement were possibly reduced by opportunities given to family members to pay a ransom, and (later in history) by 'international' agreements regulating the exchange of prisoners. Moreover, Rabinowitz's reliance on Garlan for the facts of enslavement merits some additional scrutiny. Garlan himself does not specify which historical records he used and for which historical period the figures are valid. The works mentioned in his footnotes do present a discussion of the issue, but they fail to provide the relevant statistical information. It is clear, however, from the broad chronological perspective adopted in these studies that the figure of 25 per cent is not based on evidence for the fifth century alone.[[8]] At any rate, what percentage must be considered as sufficient to produce a widespread fear that needed to be allayed in a tragic setting?

One of the pieces of evidence which Rabinowitz quotes in support of her theory is Thucydides' passage on the fate of the Athenians taken prisoner at Syracuse (7.75.7; p. 67, n. 6). Yet these tragic circumstances made such an impact precisely for the reason that the event was totally unexpected and out of keeping with the realities of an imperialistic state, signifying to Thucydides the disastrous standard of leadership after Pericles' death. Athens was accustomed to enslaving other states, in the metaphorical sense, not to seeing her citizens becoming actually enslaved. In the same footnote she argues that Solon's measures to solve the problems of debt- bondage in the sixth century may have been successful in Athens itself, but that this does not mean that the practice disappeared elsewhere, so that 'Athenian men of the fifth century could have looked around them and seen men more or less like themselves in slavery' (p. 67, n. 6). If Rabinowitz is in fact arguing that Athenian citizens bought debt-slaves from other Greek cities and kept them as their slaves, some evidence in support of this would have been welcome.

Richard Saller's contribution, although by far the shortest in the volume, contains some essential points as to which aspects of Roman culture might be most instructive in providing an idea of how women and slaves were construed. His discussion of symbols of hierarchy in the Roman household first takes up the concept of honour as the important dividing-line between the individual members of the Roman household. In this configuration married women are placed with those possessing honour, exemplified in their freedom of movement (including the right to initiate divorce independently), their legal right to own property and their entitlement to meet out punishment. For the latter aspect Saller makes two significant observations, namely that the use of the whip generally marked slave from free and that classical texts provide virtually no evidence for the beating of honourable wives. The first argument provides an eminently useful tool to study the nexus slave/free at Rome, for it bears all the signs of a fundamental distinction. As Saller has observed elsewhere, slaves in Roman comedy are mainly distinguished by their vulnerability to corporal punishment.[[9]] His second remark raises more questions, for it does not take into consideration threats of violence and a psychological climate in which these threats were validated. More importantly, his argument may not be valid, for there are some texts, some of them sufficiently neutral to discount any ideological objectives, that indicate that the beating of wives may not have been an uncommon fact of life. The meaning of the Elder Cato's dictum that men who beat their wives and children harm their most precious possession (Plut., Cato Mai., 20.2) is not unequivocal. While Saller has argued that Cato's remark illustrates that familial pietas was not an invention of the Principate,[[10]] Suzanne Dixon takes the passage to imply that physical violence was a standard element of the relations of power and hierarchy within the Roman household.[[11]] In this volume Saller's position is challenged by Patricia Clark (see below).

Saller identifies another key feature of Roman society in the distinctions between slave and free during family festivals. As he underlines in his exposition, Roman festivals are difficult to read. Lack of details on the actual proceedings, let alone on the personal experiences of those who participated (or those who were excluded), make festivals a complex set of evidence for the highlighting of gender and social hierarchies. And yet they provide some significant information. The fact that in the annual festival of the Compitalia male and female woollen dolls were hung in the cross-roads for each free member of the household, but for each slave an undifferentiated woollen ball, marks the desire to differentiate between slave and free within what was legally a cohesive unit (pp. 87f.). He further calls attention to the festival celebrated in honour of the goddess Maiestas whose purpose was to watch over the dignity of the free members of the family (p. 88) and to the Matralia which centred around the sexual jealousy of married women for their slave girls (p. 89). In a different type of Roman festival, however, the exact opposite is acted out in that the differences between honourable wives and female slaves are deliberately removed. Rome's festival calendar had a number of holidays set aside for the benefit of slaves, the most famous of which was of course the Saturnalia in December. Saller opts for the festival of the slave girls celebrated on July 7. The aetiology, reproduced in the writings of Plutarch, is an unlikely story in which slave girls replaced free matrons during an encounter with hostile Latin forces sometime after the Gallic sack, which hinges on the concepts of chastity and the indistinguishability, at least to the Latins, of slave girls in free dress and their real counterparts. This may be compared with the practice during the slave holiday on August 13 when the matron of the household washed her slaves' heads as well as her own, which Saller takes to be a reversal of the usual roles of domina and servant (p. 89; but why does she wash her own head as well?). Festivals, therefore, may provide a significant symbol of the differences between slave girls and free women, sometimes exemplified by conflict and exclusion, at other times by shared participation and the willingness of the matrona to take up, for one day only, a servile role.

The study of domestic conflict and violence in the ancient world will prove to be a very fruitful area for discussion in the years to come. In this volume Patricia Clark makes another useful contribution to the subject in a stimulating essay on hierarchy and violence in Augustine's portrayal of his mother Monnica in the Confessions. The first major point which she takes up is that of age-related behaviour and the ways in which women were socialised in their double role as wife and mistress. As a young girl Monnica was the ward of an old slave woman, whose role as guardian of social conduct is described by Augustine in purely masculine terms (p. 111). Her teachings are a reflection of the patriarchal society in which women's behaviour was embedded, combatting a range of natural excesses which reflect male anxieties (for instance, the frontal attack on the drinking of wine which is associated with the woman's role of being in charge of household supplies). The second episode shows us Monnica at a later age when she had developed a secret addiction to wine- drinking. A quarrel with a slave girl put a stop to this, when the girl called her a drunkard upon which Monnica realised her fault and at once condemned and renounced it. Clark concludes that the ancilla acts as a double catalyst, by holding up a mirror to her superior and by accelerating Monnica into the next stage of her life (p. 113). Yet there is something odd happening here which deserves more emphasis than is given to it by Clark. She argues that Monnica, 'having internalised the moral strictures of her old slave nurse,' is now taking control of herself (p. 112). I do not see any evidence for the first part of her argument. Monnica's 'coming-of-age' is the outcome of an independent development which is miraculously completed by a chance encounter with the second slave rather than the logical response to the teachings of the first. Although intended to watch over the social training of her ward, in the end the old slave is not effective in her teaching -- Monnica, after all, did become addicted to wine-drinking in spite of the lessons inculcated at a young age. In describing the variance in Monnica's behaviour at different ages Augustine subtly indicates that the socialisation of women is most effective when it is taken on voluntarily. One wonders of course if this image is not largely determined by Augustine's own ideas about the position of women.

This idea of 'instinctively knowing your place' is illustrated more clearly in his account of Monnica's marriage to Patricius and her response to her husband's violent temper. In Augustine's representation the marriage was a mixture of tension and concord, but a form of concord which implied that both partners knew their respective places. Patricius was a pagan, unfaithful to his wife and possibly an alcoholic. Monnica's strategy was to placate him when he was calm and not to resist him when he was angry. The key passage for this part of Clark's discussion describes the women of Thagaste flocking together to discuss their marital problems and to complain about their husbands beating them up. Monnica chides them for complaining so openly and suggests that they must placate or else endure. Her words persistently reflect the idea that marriage can be viewed as a form of enslavement and that by signing the marriage contract the women have submitted themselves to a life as ancillae of their husbands. Clark chooses to call this slippage and emphasizes the light-hearted way in which it was said (pp. 114; 118f.): in their marital status women were perhaps like slaves, but they were not quite the same. The Latin has veluti per iocum graviter admonens, which in Pine-Coffin's Penguin translation is rendered as 'her manner was light but her meaning serious'. Clark prefers to put full stress on per iocum ('with a light manner') and consequently downplays the importance of graviter ('in all seriousness'). However one wants to translate the passage (on p. 114 Clark gives her own translation, which is adapted from Pine-Coffin's), it is clear that in the Latin the full stress is on graviter and not on per iocum. Slippage, or metaphor, may therefore be too soft a term here, especially if Brent Shaw is right in postulating that the language of enslavement was actually in the marriage contract.[[12]]

Clark's use of parallels from earlier literature argues that on essential points Augustine is following traditional forms of discourse: 'Women are responsible for men's behavior and for changing it: by altering their own behavior, they can stop their husbands from beating them; they must placate; they should keep quiet about domestic violence and avoid making it public; and if all else fails, they must endure' (p. 115). Clark is here interested in women's strategies of mollifying their husbands, and in examples of women who maltreated and assaulted their slaves (a negative topos). Thus she sidesteps a crucial historical problem, the issue implicitly raised by Saller on p. 90, that of whether late antiquity witnessed a higher incidence of domestic violence than the classical period (it is mentioned briefly in footnote 16 on p. 127f.). The most common explanation for the discrepancy in reported cases of violence between the two periods is that of Susan Treggiari who has pointed to the importance of unilateral divorce in the pagan Roman world.[[13]] However, as Clark usefully points out, such a relationship between relaxed divorce procedures and the lack of domestic violence is problematic. This is an area where we must probe deeper. There are factors on either side which may have been equally important, such as the dwindling importance of citizenship and thus of honour, and the increased Christian focus on domestic issues prominently addressed in sermons.

The genre of Roman comedy has in recent years become an essential field for the study of various aspects regarding the Roman household. The works of Plautus and Terence form the earliest type of Latin literature that has survived more or less complete. In dealing with a mainly domestic environment comedy provides excellent illustrations of marriage patterns, the right of women to initiate divorce, and the sentimental ideal of the family. Yet the genre is problematic as a historical source.[[14]] We may surmise a historical backdrop for most of the plays, but on essential points comedy inverts developments in contemporary society rather than mirroring them in a straightforward way. This means that in order to understand how comedy may illuminate social relations in historical Rome we have to carefully weigh its evidence against external information. The problem is that in most cases it is comedy which provides our most reliable point of departure for the study of social relations. Another significant problem, indeed a point of long-standing debate, is the relationship between the Greek originals and their Roman adaptations. It is interesting to note that ancient commentators have made a careful distinction between the fabula palliata and the fabula togata. In the former slaves were allowed to deceive and dominate their masters, whereas in the latter they were not.[[15]] However this comment should be taken, it implies that the comedy in Greek dress, to which the majority of plays by Plautus and Terence belong, form a more complicated medium for the exploration of social relations.

In undertaking a study of women and slaves in Plautus' Casina Rei has set herself a daunting task, and this task is made no less difficult by her assertion that she wants to explore how 'social expectations inform the dramatic functions and characterizations of women and slaves' (p. 92). Her main contention is that the empowerment of elite women was a 'more real possibility than the empowerment of slaves' (p. 94). Rei's analysis contains some very interesting points. She argues, for example, that the sexual restraint of wives and daughters in comedy can be explained by the strict requirements of chastity for free women in Roman society (p. 95). On stage they are not only prim and proper, they are also excluded from role play and the taking on of disguises. A telling example comes from Plautus' Persa in which a daughter of free birth refuses to impersonate a prostitute out of fear that her neighbours might believe that she actually is one (p. 95; emphasis as in the original). The explanation for her unwillingness to perform a role, so Rei argues, lies in the dangerous affinities role-playing had with acting, an activity which was marginalised by Roman high society. Such a moral code did of course apply equally to free men, who, however, in Plautus, do not seem to have been hindered by it in the same way. It is perhaps more appropriate to argue that for free women to engage in role-playing represents a more serious attack on their sexual honour, arguably the only type of honour which they could possess, than for men.[[16]]

More problematic, I believe, is Rei's contention that the negative portraits of powerful wives in comedy are a direct reflection of 'the social tensions of Plautus' own cultural moment' (p. 96). The domineering wife is not part of Plautus' Greek models and may therefore comprise a uniquely Roman element. Rei, however, is frustratingly vague as to how this image may be related to historical developments. In one observation it is said to have 'coincided historically with a shift in marriage patterns that gave even greater [sic] financial power and autonomy to wives' (p. 97). Note that Rei is not claiming that historical developments have dictated the portrayal of wives in comedy, but that she is suggesting something along these lines all the same, and that she does so in a kind of language that will ensure that she can not be pinned down for making simplistic deductions ('Plautus' own cultural moment'; 'coincided historically'). Equally evasive, and in a sense completely incomprehensible, is her remark, already quoted above, that the empowerment of elite women was a more real possibility than the empowerment of slaves (p. 94). A more real possibility in terms of what? In terms of comic staging, or in terms of historical developments? Because of his/her complete lack of status, to apply the concept of empowerment to the male or female slave is something of a misnomer. 'Empowerment' can only be conceived in an informal structure of power and influence, the very foundation of the slave's prominent role in Roman comedy. The influence wielded by women, if grounded in their financial independence, may of course be a more historical phenomenon, but their powers acquire a new significance in comedy when, as is the case in the Casina, the husband is not performing his duties as a paterfamilias. Moreover, it should arouse due suspicion that wives who have gained some degree of influence over their husbands have done so in opposition to husbands who are noticeably poorer than they are. To assume, therefore, that she reflects 'the social tensions of Plautus' own cultural moment' (p. 96), is something which needs to be explored before it can be accepted, and our interpretation must surely change dramatically if women were among the audience.

Any study of women and slaves meets with serious problems, of which the lack of a perspective 'from below' is the most substantial one. The recovery of the personal voices of muted groups is a complex exercise which requires great sensitivity and ingenuity. For the study of women in the ancient world it has already been commonly accepted that such attempts are beset with too many difficulties to be rewarding. For the study of slaves the prospects are even less promising, although comparative studies may be able to throw some light on certain matters. It is of crucial importance to realise that no ancient source refers to prose or poetry written by slaves, even though some of the best-known intellectuals were of servile origins.[[17]] What are commonly understood as texts in which the experience of slavery acquires a personal voice have all been written by ex-slaves. The distinction between slave and ex-slave is essential for after manumission the slave was eager to build a new identity out of a spectrum of ideas which had no connotations with a servile existence, a process which modern scholars have coined imitatio domini. The extant writings of ex-slaves (Epictetus; Aesop) betray few significant traces of a former slave psyche, and any effort which attempts to uncover them fails from a lack of a clearly defined understanding of what the slave psyche entails.[[18]]

Some of the complexities involved in the recovery of the original slave experience are demonstrated in this volume by Shane Butler who discusses a slave's act of self-castration recorded by the obscure writer Julius Obsequens in his Prodigiorum Liber. The text itself is brief: servus Q. Servilii Caepionis Matri Idaeae se praecidit et trans mare exportatus ne umquam Romae reverteretur. ('A slave of Q. Servilius Caepio cut himself off in front in honour of the Idaean Mother and was shipped away across the sea, that he might never return to Rome', translation by Butler on p. 239). Butler first makes use of traditional empirical methods of investigation to make sense of this bizarre act. The event itself took place in 101 BC and Butler tantalizingly connects it with the historical context of the second Sicilian Slave war. He offers the suggestion that the slave's act was one of solidarity, although he admits that this can only be a conjecture. The slave is said to have belonged to Q. Servilius Caepio, most probably the consul of 106, who, in 103, was convicted of maiestas (a term which covers various forms of treason) by the tribune Gaius Norbanus, after which his property was confiscated; he escaped imprisonment by fleeing to Smyrna. This implies that in the year in which the castration took place the slave himself had already been allocated to a new owner. The most likely solution is that he became a public slave serving in the sanctuary of the Idaean Mother. Butler argues that some words have fallen out in between 'Caepionis' and 'Matri', and that the dative case should be emended to a genitive, suggesting a location ('in the sanctuary of . . . ', or something similar) rather than be taken for a special relationship with the goddess (p. 242). The injunction ne umquam Romae reverteretur ('that he might never return to Rome'), Butler argues, concerns Caepio's position rather than the ultimate fate of his former possession. Butler reconstructs a plausible scenario in which Caepio, with the aid of friends in Rome, was orchestrating a possible return. The erratic behaviour of one of his former slaves was regarded by the authorities as a sign that such a return should be prevented at all costs.

Butler employs a wide range of not inconsiderable skills to contextualise a Latin sentence of some 15- odd words, but in the end he has to admit that we cannot come closer to an understanding of the slave's emotions at the time (p. 243). I find this unsurprising. The political and ideological framework that has been established suggests that the slave's thoughts which led to his action were irrelevant to those who ultimately recorded it in its present form. By labelling it a prodigy the authorities have attempted to give meaning to an otherwise inexplicable phenomenon. Butler's negative findings are not without interest, exemplifying how the interests of a minority triumphed over the silenced intentions of an anonymous slave, whose behaviour can only be explained (by them) in terms of his propensity to be his master's voice. Yet I am slightly taken aback by the amount of misdirected energy in search of evidence which this text, because of its brevity, its truncated Latin and the genre to which it belongs, is unlikely to deliver.[[19]] Butler's approach bears some similarity with the one adopted by David Cohen in a recent discussion of Roman adultery.[[20]] Here the behaviour of Vistilia, a senatorial woman who registered herself as a prostitute, is at issue and Tacitus' suggestion of depravity makes it equally difficult to find clear evidence of the original motivation for this behaviour. However, where Cohen is able to frame the Vistilia episode in a broader socio-historical context, Butler's discussion remains fixed on the behaviour of an individual slave. He does not move beyond a couple of stereotypical observations on lower-class participation in the cult of Cybele and the dimension of sexuality is too insubstantial to provide a convincing framework.

A combination of the study of gender and slavery is an intrinsically rewarding exercise, and one of the areas which greatly benefits is the position of the female slave, arguably the individual which ranked lowest on both gender and status in Greco-Roman society. The amount of attention devoted to female slaves is in line with recent developments in the study of other slave societies, a fact which is used to good effect in most of the essays.[[21]] The variety of typologies, however, is at times confusing, for throughout the volume there is no clear distinction between the specific roles which female slaves might perform and how they are perceived in literature. The approach which is adopted by most of the contributors to this volume gains a fair number of insights into how in literature social categories are the subject of a continuous process of redefinition. From this perspective the volume offers a good showcase of how awareness of this process can further our understanding of how most ancient literature served the ideological purposes of a small socio-cultural elite. The message carries conviction and is well illustrated. Yet there is a certain limitation to its explanatory value and some of the authors carry their arguments too far, especially when the theme is deemed to carry the whole structure of the text. Readers will find it remarkable to discover in certain chapters almost no references to alternative explanations of the main points or indeed the thrust of the entire text.[[22]]

Of the dual perspectives of cultural representations and social realities which the volume aims to discuss, it is the former which is best served, but how far is this concept going to take us towards a better understanding of the latter? Put somewhat differently, is the relationship a tied one so that one might be able to view shifts in cultural representations as keys to historical changes, or do they lead separate existences? If neither of the two options offered here can be correct for each individual case, the relationship as such needs to be made the subject of a clear analysis, or at least a number of preliminary observations, for otherwise, as happens for instance in some of the chapters which have been reviewed in more detail, observations of an ideological nature found in ancient texts will be taken as more or less the equivalent of historical comments. This matter is addressed in more general terms by Richard Saller. He argues that literature and rhetoric constituted only two among the many spheres of symbolic discourse that contributed to the construction of Roman categories and hierarchies of status and gender. He states that it is 'the task of social historians to explore the other, non-literary, spheres, both verbal and non-verbal, and to ask how they reinforced or stood in tension with the literature written by elite men' (p. 85). Such vital theoretical considerations are too frequently absent from the other discussions.[[23]]

NOTES

[[1]] For a recent discussion of the scarcity of female authors in the Roman world, cf. Thomas N. Habinek, The Politics of Latin Literature: Writing, Identity, and Empire in Ancient Rome (Princeton 1998) 122-37, and the older literature on the topic on 209f.

[[2]] Cf. eg W. Spickermann, Mulieres ex voto: Untersuchungen zur Götterverehrung von Frauen im römischen Gallien, Germanien und Rätien (1.-3. Jahrhundert n. Chr.) (Bochum 1994). The classic study on slave mentality and religion is F. Bömer, Die Religion der Sklaven: Untersuchungen über die Religion der Sklaven in Griechenland und Rom, four volumes (Wiesbaden 1957-1963).

[[3]] It would be of some significance to know whether these women wrote while they were still enslaved or after the abolition of slavery, a point which is not discussed by McCoskey.

[[4]] Cf. Victoria Wohl, Intimate Commerce: Exchange, Gender, and Subjectivity in Greek Tragedy (Austin 1998) esp. 111-14, where Cassandra is viewed as the main focaliser for all the male suffering in the play. Wohl also underlines her use of gendered language in describing Clytemnestra's crimes. The best quote can be found at 113 where Wohl analyses the connections between Iphigeneia and Cassandra: '[Cassandra] is the fantasy of the sacrifice scene come true, a woman loyal to men'.

[[5]] Cf. N.T. Croally, Euripidean Polemic: The Trojan Women and the Function of Tragedy (Cambridge 1994; not mentioned by Rabinowitz), esp. 97-103, for a less strictured analysis of the polarities of slave and free.

[[6]] Cf. N.R.E. Fisher, Slavery in Classical Greece (London 1993) 71f., who discusses Lysias, On the Olive Stump 7.34f. and Plato, Republic 578d-9b.

[[7]] Y. Garlan (tr. Janet Lloyd), Slavery in Ancient Greece (Ithaca and London 1988) 48.

[[8]] The main works referred to by Garlan are: Hans Volkmann, Die Massenversklavungen der Einwohner eroberter Städte in der hellenistisch-römischen Zeit (Wiesbaden 1961); Pierre Ducrey, Le traitement des prisonniers de guerre dans la Grèce antique, des origines à la conquête romaine (Paris 1968). Anne Bielman, '*LU/TRA, prisonniers et affranchis', MH 46 (1989) 25-41, should be added to the list in Garlan.

[[9]] Richard P. Saller, Patriarchy, property and death in the Roman family (Cambridge 1994) 137f.

[[10]] Saller [9] 6, n. 17.

[[11]] Suzanne Dixon, 'Conflict in the Roman Family', in Beryl Rawson & P. C. Weaver (eds.), The Roman Family in Italy (Oxford 1997) 150, n. 3.

[[12]] Brent D. Shaw, 'The Family in Late Antiquity: the experience of Augustine', Past and Present 115 (1987) 3-51, esp. 28, n. 104; 35, n. 139; 35f. 13; Susan Treggiari, Roman Marriage: Iusti Coniuges from the Time of Cicero to the Time of Ulpian (Oxford 1991) 430f.

[[14]] Cf., e.g., Richard P. Saller, 'The Social Dynamics of Consent to Marriage and Sexual relations: The Evidence of Roman Comedy', in Angeliki E. Laiou (ed.), Consent and Coercion to Sex and Marriage in Ancient and Medieval Societies (Washington, D.C. 1993) 83-109.

[[15]] Rei does mention this observation in one of her footnotes (p. 107, n. 59), without, however, drawing it into the main discussion.

[[16]] To the Roman mind female pantomimes were easily associated with prostitutes, cf. H. Leppin, Histrionen: Untersuchungen zur sozialen Stellung von Bühnekünstlern im Westen des römischen Reiches zur Zeit der Republik und des Principats (Bonn 1992) 137.

[[17]] All the more interesting is the tantalizing piece of information that in the second century AD a certain Hermippus of Berytus composed a treatise entitled On Slaves Famous in the Cultural Domains, cf. S. Mazzarino, Il pensiero storico classico, II 2 (1966) 131-99. For criticism of Mazzarino's views, cf. M.I. Finley, Ancient Slavery and Modern Ideology (London 1980) 117; 177, n. 96.

[[18]] For Epictetus see now J.P. Hershbell, 'Epictetus: a freedman on slavery', AncSoc 26 (1995) 185-205. Hershbell is far too optimistic about the degree to which the Diatribes provide insights into the psychological effects of slavery. Epictetus' writings may contain more observations on slavery than can be found in any other philosopher of the imperial period, but they do not necessarily betray a higher sensitivity to the plight of slaves.

[[19]] Butler is apparently unaware of a more successful attempt at reconstructing slave perceptions by Keith Hopkins, 'Novel Evidence for Roman Slavery', Past and Present 138 (1993) 3- 27.

[[20]] David Cohen, 'The Augustan Law on Adultery: The Social and Cultural Context', in: David I. Kertzer & Richard P. Saller (eds.), The Family in Italy: from Antiquity to the Present (New Haven and London 1991) 109-27.

[[21]] A useful work on this subject, which is not included in the bibliography, is Barbara Bush, Slave Women In Caribbean Society: 1650-1838 (Kingston 1990).

[[22]] Since I have referred to important omissions in previous footnotes, one example may suffice here. In her discussion of Plautus' Casina Annalisa Rei fails to make mention of three important articles which explicitly deal with the problems the play poses and omits mention of a standard work on slavery in Republican Rome which covers the slave characters in Plautus and Terence extensively: Jean Christian Dumont, Servus: Rome et l'esclavage sous la république (Rome 1987); W.E. Forehand, 'Plautus' Casina: An Explication', Arethusa 6 (1973) 233-56; Jane M. Cody, 'The Senex Amator in Plautus' Casina', Hermes 104 (1976) 453-76; Bronwyn Williams, 'Games People Play: Metatheatre as Performance Criticism in Plautus' Casina', Ramus 22 (1993) 33-60.

[[23]] For those interested in metaphors of slavery, cf. Peter Garnsey, Ideas of slavery from Aristotle to Augustine (Cambridge 1996), a work which to my surprise appears not to have been consulted by any of the scholars contributing to this volume. This may be due to the fact that it appeared too late to be used with profit, especially since the chapters have their origin in a meeting of the Women's Classical Caucus, held at the APA meeting in Atlanta in December 1994. Yet the bibliography contains many items published after this date. On pp. 220-35 Garnsey has a useful analysis of Christian texts in which sons and slaves are compared, and his comments are interesting, not only because they concern a process comparable to the metaphorical usages of slavery discussed in Murnaghan and Joshel, but also for the theoretical implications that he draws, cf. esp. pp. 234f.: 'It is worth asking whether, for the Church Fathers, slavery and sonship had two more or less independent existences, one metaphorical, in the land of theology, the other physical, in Greco-Roman society. The two worlds seem to me to have intersected surprisingly little'.