The Canonised Classics

On the surface, canons appear to be an almost inescapable function of institutionalised knowledge – a hierarchy of authors, texts and methodologies that structure both our curricula and the profession in decisions on appointments and promotions, conferences and grants. As John Guillory writes: ‘the canonicity of works is … another name for their institutional mode of reception and reproduction’.Footnote ¹ Canons, so the argument goes, are practical tools for creating and maintaining learning communities and for negotiating belonging to groups (academic, social, professional, etc.). While acknowledging the need to change the boundaries of the canon to include under-represented texts or readers, proponents of the pragmatic approach to canonical formations emphasise what they see as the sheer inevitability of the canon. As the argument goes, a canon, however negotiated, is indispensable for teaching and research: how else but through the negotiations of canonical formations in syllabi, reading lists, conferences and funding bodies is one to carry out the paedagogical and professional missions with which we are entrusted as scholars and teachers?

It is important to emphasise from the start that this pragmatic argument betrays defensiveness. Whether explicitly or not, such pragmatic arguments are often made in response to critiques from marginalised groups: students and scholars who are critiquing the exclusion of racial, ethnic and religious minorities from academia and are seeking to create canons, reading lists and institutions that are representative of the diversity of society. The rhetorical context in which these ‘canons of practicality’ are often evoked should give pause for thought even to the most committed of pragmatists.Footnote ² What critics of the canon object to is not just the content of canonical formations – canons that are all too often too male, too white and too western-centric – but also the ways in which the supposed aesthetic superiority of Homer, Virgil, etc., which is repeatedly invoked to justify the inclusion of these works into the canon and the exclusion of others, erases the process through which canonical texts have been used to defend and rationalise violence against racial, religious and other minorities and against women.Footnote ³ Such a history of reception, canon critics argue, is one that should be represented in paedagogy, and may even warrant exclusion from the canon or at the very least a redefinition of a given text’s role within it. Thus canon critics do not simply want to redesign the canon out of a self-promoting agenda as charged by opponents, who often implicitly present themselves as promoters of the practical or objective defenders of great literature. Rather, they are inviting much-needed reflection on the mechanisms underpinning canonical formations, beginning with the nexus of aesthetics, ideology and power through which the canon carries out its ‘practical’ mission.Footnote ⁴

An analysis of the canon must thus begin by reckoning with the work that canons have traditionally performed in literary studies and the role that Graeco-Roman antiquity has played in the formation of the broader discourse of canon. Even the most ardent proponents of the canons of practicality might concede that it would be reductive to see the relationship of canon and the classics of Graeco-Roman antiquity merely as the result of the mechanics of academia.Footnote ⁵ For one, there is practically an equation between classical antiquity and the canon: by its very name, Classics has traditionally legitimised itself as a field through its access to objects that claim to have intrinsic trans-historical value. Friedrich August Wolf’s Prolegomena to Homer is hailed as the beginning of the disciplinary field of Classics to the exclusion of, for example, Christian texts, Greek Judaism and the languages of imperial peripheries.Footnote ⁶ If Classics is in effect co-extensive with Homer, the author who most embodies the canon, it is inevitable that it should be considered a ‘highly canonised’ field.Footnote ⁷

More to the point, as we have seen above, definitions of the classical have been traditionally invested in claims of the universal superiority of the Graeco-Roman tradition and therefore implicitly of western culture. In the words of Hans-Georg Gadamer, the classical is ‘a kind of timeless present that is contemporaneous with every other present’ (Reference GadamerGadamer 2004: 299). Here Classics is deeply complicit in the formation of Euro-centrism, western-centrism and other elitist ideologies, which use the Graeco-Roman past as the underpinning for claims of trans-cultural and trans-historical superiority. It has been repeatedly observed that the notion of the classic is inextricably connected to that of empire, as for example in T. S. Eliot’s terse formulation from a lecture delivered to the Virgil Society at the height of World War II in October 1944: the Aeneid is the classic par excellence because ‘Aeneas is the symbol of Rome; and, as Aeneas is to Rome, so is ancient Rome to Europe. Thus Virgil acquires the centrality of the unique classic; he is at the centre of European civilization, in a position which no other poet can share or usurp’ (Reference EliotEliot 1974: 28–9).

The co-extensiveness of the classical with timeless trans-cultural values can take several forms: a ‘quasi-Darwinian’ claim that the value of the classics is tied to and evidenced by their continual survival. In this reading, ‘the canon is … a gauge of vitality’ (Reference BloomBloom 1995: 39), ‘the classic … defines itself by surviving’ (Reference CoetzeeCoetzee 2011: 20) or, to put it differently, it is democratically defined by the ‘pleasure’ it gives to successive reading communities.Footnote ⁸ Exceptionalist claims can also be found lurking in arguments about the value of the classical for understanding European culture or for placing western civilisations in dialogue with non-western ones.Footnote ⁹ The pitfalls of these paradigms are not hard to spot. On the one hand, there is a danger of constructing the canon as a universal mode and thus imposing culturally alien categories on non-western traditions.Footnote ¹⁰ On the other, as Emily Apter has reminded us, the methodology behind ‘global literature’ conjures practices of ‘collectivism’ and ‘curatorial salvage’ and other markedly colonial modes through which literature is cast as ‘property’, worth collecting, preserving and assessing in its ‘value’.Footnote ¹¹ Apter puts her finger on the ways in which the classical can play a role in institutional forms of knowledge that are deeply entwined with western notions of personhood and a neoliberal emphasis on the transference of goods and market value. Finally, critics have attacked from different disciplinary angles the very narrative of a cohesive western canon as inherently fallacious and blind to the porousness of east–west categories.Footnote ¹² We could recall Amitav Ghosh’s attack on the myth of a cogent European Novel in the essay on his grandfather’s library, which is in itself a deep critique of the discourse of the canon. Ghosh denounces the Novel as a genre both deeply rooted in a local and parochial European identity and yet intrinsically dependent from the start on imperial dislocation for its existence.Footnote ¹³

Other critics such as Pierre Bourdieu and Barbara Herrnstein Smith have focused our attention on the pervasiveness of the language of value in literary analysis and the meaning of this productive intersection of economics and literature, investigating the kind of ‘cultural capital’ that the canon represents and how it functions in a system of exchanges within academia.Footnote ¹⁴ Central to this approach to the canon as an institutional mechanism is the critique of the contingent nature of literary judgements: the value of the text is not a property residing in the object of study but rather a positive effect arising from readers’ engagement with it. In turn, those engagements are pre-shaped and pre-determined by the institutional context in which they take place.

For whom, with whom and where we are judging and reading becomes key to understanding the very functioning of the canon as a tool for hegemonic control.Footnote ¹⁵ Indeed, work by Ankhi Mukherjee on the role of the canon in post-colonial writing has unearthed the ways in which ‘the question and concept of the classic is … always that of the outsider’.Footnote ¹⁶ Commenting on the essays by the same title – ‘What is a Classic?’ – by T. S. Eliot and J. M. Coetzee, respectively written in war-torn London in 1944 and delivered in 1991, a year after the end of the Apartheid regime, Mukherjee delves deep into the ways in which both writers use the classics to explore and tackle their own peripheral status in relation to Europe (in the case of Eliot, as an American living in the UK) and the global literary stage (in the case of South-African Coetzee). Recent work on constructions of the classical in Graeco-Roman antiquity has placed a similar emphasis, for example, on the ways in which visions of the Greek classical past are shaped by the geographical and temporal liminality of imperial Greek writers of the Second Sophistic and also on how this ancient construction of Hellenism in writers such as Plutarch and Aelius Aristides has been privileged in the service of a romanticised and anachronistic attachment to a pure version of Hellenism in nineteenth-century scholarship.Footnote ¹⁷ A comparable critique of the scholarly investment in the construction of the classical past has come from the field of late antiquity.Footnote ¹⁸ Here in putting forward a vision of late antiquity as a period of extreme dynamism and vitality, even while retaining a label that explicitly relegates the period to a liminal space relative to classical antiquity, scholars have pointed to the heuristic value of the Gibbonian construct of decline and fall not for late antiquity but rather for Enlightenment history.Footnote ¹⁹

It takes not only imagination and risk-taking to question the inevitability of the canon as an enshrined institutional and professional practice but also an intimate knowledge of the history of higher education, a high bar to clear for most classicists, including this author. The classics, including Homer and Virgil and other Graeco-Roman texts, have been the bedrock of the modern liberal arts curriculum from its earliest inception. One of the most significant documents in the history of American Higher Education, the Yale Report of 1828, contains one of the first arguments for a liberal education, as opposed to a professional one, defined as ‘such a course of discipline in the arts and sciences, as is best calculated, at the same time, both to strengthen and enlarge the faculties of the mind, and to familiarise it with the leading principles of the great objects of human investigation and knowledge’ (p. 30).Footnote ²⁰ The best course of study is one which brings into exercise all the important mental faculties and is therefore both propaedeutic to further specialised study in the professional school and conducive to engaged participation in society by all members to whom such education might be accessible. It is important to note, however, that the case for this liberal arts curriculum inclusive of science, mathematics and what we may now term ‘Humanities’ is advanced in the context of the question as to whether Greek and Latin should be retained as part of the college instruction or instead eliminated because irrelevant to most professions. In some ways, though 200 years old, the report is strikingly applicable. The slippage between practicality and desirability is already at work in this early document: according to the authors since ‘the literature of every country in Europe is founded more or less on classical literature … if scholars are prepared to act in the literary world as it in fact exists, classical literature, from considerations purely practical [emphasis added], should form an important part of their early discipline’ (p. 34). The authors make a limited claim as to the value of Greek and Latin for those going into law, medicine and theology. While Greek is especially valuable to access scripture in its ‘original simplicity and purity’, Latin is framed as essential for those studying modern European languages. Yet on closer inspection, the practicality of the liberal arts canon is but one part of a larger and more serious claim about the ability of the Greek and Roman writers to ‘form the taste, and to discipline the mind, both in thought and diction, to the relish of what is elevated, chaste, and simple’ (p. 35).

The political nature of this discourse, both in its original context and in its subsequent instantiations, cannot be understated:Footnote ²¹ for as scholars have demonstrated, this rhetoric went hand in hand with an appropriation of the classical past to justify modern structures of colonialism and slavery.Footnote ²² To give one contemporary example, John Calhoun, the American slavery advocate, statesman, vice-president of the United States and Yale graduate, is rumoured to have asserted that philological competence in Greek was an indicator of humanity, and one in which black people were incapable of partaking.Footnote ²³ For sure these constructs were contested already in the nineteenth century:Footnote ²⁴ African American scholars, educators and intellectuals not only actively combated these appropriations by using classical texts to challenge racist ideologies but also co-opted Graeco-Roman civilisation in the development of their nascent programme for African American culture.Footnote ²⁵ Nevertheless, even as the authors of the Yale Report wrestle with subtlety with transatlantic differences in the teaching of Greek and Latin at University, there is no question that the goal for which this curriculum is designed is the formation of a white male who claims European heritage as his own.

Thus, though the reconfiguration and supplementation of the curriculum in today’s Anglo-American classrooms may on some level satisfy the present demand for greater inclusivity, it is important to acknowledge that the place of classical works in the liberal arts canon rests on a mixed legacy. To give one other example, in the US, debates about opening up the canon have clustered around the redefinition of the curricula for ‘Great Books’ courses in the context of liberal arts education in places such as Reed College, Columbia, Chicago and Yale. Demands for Great Books programmes, such as Columbia’s Core Curriculum as well as similar programmes at Chicago, Berkeley, University of Virginia and later at Yale, arose in the aftermath of World War I in an effort to strengthen democracy and to avoid the mistakes that led to near annihilation in the first world conflict.Footnote ²⁶ Originally developed as war issues training and accompanied soon by an Honours course on Great Books taught by John Erskine, ‘An introduction to contemporary civilization’ was designed to teach students among other things ‘how to produce cheap goods without sacrificing human nature; how to achieve political and legal forms that are at once flexible and stable; how to eliminate human and material waste of any kind; how to preserve national integrity and still enjoy the benefits of international organization’.Footnote ²⁷

On the one hand, the founders of these programmes, men like John Erskine at Columbia, Charles W. Eliot at Harvard and Mortimer Adler at Chicago, were driven by the idea of democratising both access to the classics and their interpretation – the innovation of Great Books programmes resting as much on their curricula as on the Socratic method employed in the classroom.Footnote ²⁸ This attempt to ‘democratise’ the classics was partly felt to be an answer to the development of electives in university curricula and the increased ‘professionalisation’ of higher education.Footnote ²⁹ It is in response to these changes in higher education that a common curriculum, based on a set of texts once considered the bastions of social elites, was developed to educate this student body made increasingly of middle-class men from first- or second-generation European immigrant families. Here the classical past, specifically inflected as Greek, provides a trait d’union in the melting pot of the nascent American democracy. As Erskine writes in his essay ‘The moral obligation to be intelligent’:

On the other hand, the emphasis on the classics’ role in catalysing social and ethnic assimilation of recent immigrants and inculcating a new democratic moral order is not just built on a skewed and idealised image of the classical past but, as James Turner has argued, also can be seen to mask a Christian, and specifically Protestant, agenda by endowing the teaching of literature with ‘the promise of moral formation once shouldered by Protestantism’.Footnote ³⁰ Still, perusing Erskine’s manifesto on reading Great Books one is struck by the vibrant modernity of many of his claims: in carving a space for the non-scholar to read Great Books, Erskine argues that the historical method of reading the classics does not deserve the credit it gets:

This call to examine Virgil as a book for the present has its roots in a radically modern concept of reading as a dynamic process, through which change is exerted on the past:

At least two competing models of canonicity are seemingly at work. In one, the Graeco-Roman past is presented as embodying timeless values accessible through historical study. As in the case of the Yale Report of 1828, the classical canon thus conceived is a practical tool through which to read the present and shape the future. The other, represented in John Erkine’s Great Books school, approaches the classics and the canon to which they belong more consciously as a displacement of the self, ‘a magic surface, in which [the audience] can see themselves more clearly than elsewhere’ (Reference ErskineErskine 1935: 23). The canon is here envisaged as a currency, a tool to make out who we are in the present moment, as opposed to a vehicle for the transmission of quantum values. It is and should be changeable not just in scope but also in definition, with works being swapped in and out as well as being radically redefined through reading. In many ways, however, both of these versions of the classical canon activate, self-consciously or not, a deep sense of intimacy with the classical past, whether through the promise of repetition or through the distorting effect of the mirror.

Thus the creation of a liberal arts canon, spanning the Graeco-Roman classics and European literature and thought all the way to the nineteenth century, implies a politically charged claim of continuity between the present and the classical past. Such a claim is deeply connected with nineteenth-century discourses of nationalism in all their sinister ramifications (e.g. anti-Semitism, Euro-centrism, Orientalism).Footnote ³¹ To give just one example of this complex legacy, American Great Books activists, including John Erskine himself, were deeply influenced by the programme for liberal education founded on the classics developed by Matthew Arnold, the Victorian poet and critic.Footnote ³² Arnold argued that culture was not one based on ‘a smattering of Greek and Latin’ or ‘scientific passion’ but rather that it was to be defined as ‘the study of perfection’:

In ‘The function of criticism at the present time’, an essay published in 1865 after Arnold delivered his famous lectures ‘On translating Homer’, the critic asks ‘There is so much inviting us! What are we to take? What will nourish us in growth towards perfection?’Footnote ³³ The answer lies in selecting ‘the best that is known and thought in the world’ [emphasis added].Footnote ³⁴ While Arnold invites every critic to possess ‘one great literature, at least, besides his own’ (p. 82), his model of great literature is in practice centred on the revival of a cogent European culture founded on distinct national identities but sharing a common heritage in antiquity:

Moreover, as Miriam Leonard has shown, Arnold’s socially liberal programme of education through ‘great literature’ is in itself highly fraught: Arnold’s famous call, also advanced in his Reference Arnold1869 Culture and Anarchy, to strike a balance between Hebraism and Hellenism, which Arnold viewed as the two idealised cultural forces animating world history, is both a corrective to contemporary discourses of philhellenism by French Semitics scholar Ernest Renan, among others, and an extension of it, being in no way exempt from the odious anti-Semitic rhetoric which animates it.Footnote ³⁵

In highlighting the complex and at times deeply troubled history of the instructional contexts in which the study of Greek and Roman Great Books is embedded the point should not be to seek out alternative ‘bias-free’ textual canons but rather to emphasise the political nature of these paedagogical projects and the claims to greatness therein advanced.Footnote ³⁶ Furthermore, as we construct a history of the Latin canon, we must reckon with the role of the Graeco-Roman past in the construction of modern nostalgic canonical formations. Canons may well be practical tools but they are ideological nonetheless.

‘Opening Up the Latin Canon’

When canonicity and the canon are framed as the ‘price of doing business’, the focus shifts instead to the process of negotiating and ‘opening’ up the canon to the under-represented or marginal. When we consider the canon first and foremost as a function of institutional practice – a hierarchical list of authors, texts and methodologies deemed for one reason or another worthier of being taught, studied and preserved – there is no question that in the past three decades, the core of Latin literature has significantly widened with a boom in areas such as post-Virgilian epic, technical literature and miscellaneous collections (e.g. Columella, Pliny the Elder, Vitruvius, Frontinus, Valerius Maximus, Aulus Gellius), pseudepigrapha (e.g. Appendix Vergiliana) and late antique literature to name a few. New theoretical models and emerging methodologies have aided this expansion: some of the recent shifts that have fuelled – and in turn have been fuelled by – the expansion of the canonical core of Latin texts include, for example, the erosion of traditional boundaries between literature and history and between text and material culture, and the booming field of reception studies which has significantly expanded the historical and geographical scope of the classical scholar. The rise of a mass market for translations and textbooks, stimulated by the need to teach Greek and Latin texts in courses taught in translation to a broader set of undergraduate audiences, has further shaped the curriculum to allow for the inclusion of corpora, such as papyri, graffiti and inscriptions, previously difficult to access.Footnote ³⁷ The extent to which the digital age will affect the curricular canon is still in question: on the one hand, the internet has made more material more easily available through searchable databases and yet the price of digital subscriptions and the practicalities of the classroom may in the end push more readers towards older translations out of copyright and therefore freely available on the web.Footnote ³⁸ In turn, the older the material, the more likely it is to be alienating in its deployment of racist and sexist vocabulary and in its blatant misappropriation of the classical past.Footnote ³⁹

Most recent discussions of the canon have focused on drawing and re-drawing the boundaries of Latin and Greek literature to include the writings and perspectives of under-represented or marginalised social, racial and gender groups, a re-kindling of the canon wars of the 1980s and 1990s.Footnote ⁴⁰ It is important to consider both the politics of this ‘opening up’ of the canon and how such a gesture plays out with respect to the specific field of Roman literary studies and more generally of Classics. In this section, we will explore the obvious paradox at play whenever we critique the canon by calling for a change in its internal composition without calling into question its validity as a construct. As Spivak notes in the quote which opens the section, these ‘alternative canon-formations’ are nothing but ‘variations upon an old standard’. If, as John Guillory writes, ‘the point is not to make judgment disappear but to reform the conditions of its practice’, what are the limits of this logic of ‘substitutability’ not just for the study of the marginal but also for the discipline?Footnote ⁴¹

Work that seeks to recover the experiences of marginalised groups and the gender and race ideologies promoted by Graeco-Roman texts has laid important ground for framing traditionally peripheral constituencies as part and parcel of dominant texts and ideologies.Footnote ⁴² This body of research has emphasised that Graeco-Roman antiquity is in many ways in a relationship of deep alterity from modern democratic ideals of inclusivity, fairness and democracy and such an acknowledgement is surely a salutary corrective to the traditional view of Greece and Rome as the cradle of all that is good about western culture. Furthermore, the emphasis on the linguistic, racial, religious and cultural plurality of the ancient world not only provides a richer and more faithful picture of Graeco-Roman antiquity but defuses the selective appropriation of Rome and Athens as founding sites of white European culture.Footnote ⁴³ Still, there are obvious limits to this approach to ‘opening up’ the canon in the field of antiquity, in which women, slaves and other minorities had restricted access to literacy. Despite the exciting volume of new work on categories and voices traditionally neglected or altogether excluded from scholarly discourse, as scholars of antiquity we must be careful to guard against what David Damrosch has called the ‘insistent presentism’ of postmodernity, resisting the ways in which the impetus to ‘modernise’ the canon excludes periods and genres in which the voices that seek representation are less present or less clearly visible.Footnote ⁴⁴

Relatedly, the incorporation of post-classical reception in the classical curriculum is another major engine of democratisation of the canon.Footnote ⁴⁵ Through the study of classical reception in its global dimension across time and space, the relevance of the Graeco-Roman world to different and mutually connected cultural traditions is explored in all its complexity. This emphasis on the plurality of voices that have a stake in the classical tradition is intellectually enriching as well as reparative at a time when issues of access, relevance and racism are increasingly being raised in the field.Footnote ⁴⁶ While this expanded and more open-ended vision of the classical canon has met with resistance in some quarters with a growing sense of discomfort both at what is being perceived as a ‘de-skilling’ and loss of disciplinary competence and at the erosion of the exemplarity of the classical, other work has focused on nimble, capacious and pragmatic paradigms for constructing canons that accommodate antiquity and its reception.Footnote ⁴⁷

These expanded, revised, opened or upended canons raise fundamental questions about the mechanics of artistic representation and the role of communities in hermeneutical dynamics. First, the notion that, for example, a male-authored text is bound to transmit a male-centred world-view rigidly restricts the meaning of texts to the intentions of their author, obscuring the ways in which communities do not ‘find’ or ‘stumble upon’ meaning but actively construct it in the process of reception. Moreover, if recent work on identity has taught us anything, it is that identities are far from monolithic self-evident entities: they are rather intersectional (Crenshaw), performed (Butler) or even ‘idealised imaginings’ (Pratt).Footnote ⁴⁸ When Stanley Fish poignantly asks whether there is ‘a text in this class’, we would be well advised to consider this as a question not just about the unity and comprehensibility of literary texts but also about the possibility of arriving at a definition of stable and homogenous identities based on ‘shared values’ of any one community.Footnote ⁴⁹

Thus the idea of opening up the canon raises critical questions about the construction of identity, community and readership. When texts are viewed not as static conveyers of meaning but instead as objects that come into being only in the process of interpretation, the question of whose identity, history and values are represented by or through the object will necessarily involve a complex consideration of the identities not just of makers, creators and authors but also of contemporary readers and audiences, as they collaborate to shape its culturally and historically situated interpretation. In this perspective, the diversifying of the canon has to be understood as a dialogic project involving both the identities of readers, students and scholars and the subject matter of their study.Footnote ⁵⁰ Opening up the canon is a dynamic and self-reinforcing process and one which involves both readers that embody difference (social, racial, gender, etc.) accessing and studying an expanded and evolving canon and texts (peripheral, post-classical, marginal, etc.) that embody difference being ‘read into’ the canon by an increasingly diverse readership. As Lorna Hardwick shrewdly points out with reference to the question of whether Classics has undergone a ‘democratic turn’, participation is a key measure of democratisation in ensuring that Classics is not an exclusive practice and one historically co-opted to justify conservative, reactionary or hateful programmes.Footnote ⁵¹ Yet, as Hardwick argues, access is measured not just quantitatively but also qualitatively as the process of ‘re-vision’ (Reference RichRich 1972), identifying paths not taken and thus revealing ‘hitherto unrecognised possibilities’ (Reference WolfWolf 1988: 270–1) that are activated whenever the subjectivity of new readers comes into play with the work’s crevices and inconsistencies.Footnote ⁵² Opening up the canon begins by taking seriously the inherent fragmentation of texts and by acting out our desire to cure the loss through a reading that is transformative for both the subject and the object.Footnote ⁵³

If approached from this angle, the process of opening up the canon involves a deep methodological shift rather than a process of substitution, an upending of priorities or, worse still, a widening that feels unmanageable. But how does this methodological shift affect approaches to specific texts and genres, especially as they pertain to Latin literature? First, the strong traction of the concept of canonicity within the field of Classics determines not just a hierarchy of texts and methodologies worthier of being taught and researched but also informs the very approach to so-called marginal texts. As a scholar of ‘minor’ Roman poetry, in my own research on Latin pseudepigrapha, I have had to overcome an intellectual impasse. Scholarship on Roman fakes has generally consisted of studies of individual texts aimed at proving or disproving the conventional attribution and dating: are any or all the poems of the Catalepton by Virgil? Is the Laus Pisonis Neronian or Flavian? And is the writer the young Lucan? Is the writer of the Consolatio ad Liviam also the author of the Elegiae in Maecenatem? Scholarly engagement with these texts traditionally stops at whatever provisional answer one has reached on these endless and endlessly unanswerable questions, and little or no thought is given to the dynamics that gave rise to these texts.Footnote ⁵⁴ The anonymity of these works is one of the factors that have positively discouraged any form of literary analysis, tied as this has traditionally been to the process of constructing and reconstructing authorial identity.Footnote ⁵⁵ Yet, no less influential a factor in discouraging any interest in the cultural roots of anonymous and spurious works is the self-proclaimed role of the philologist as a defender of the canon. Here the canon cultivates a ‘policing mentality’ in which a select group of works considered valuable has to be kept free from unwanted impostors, while little or no interest is given to the literary dynamics of the works rejected as spurious. One is reminded of Antony Grafton’s compelling image of antiquity as a train ‘in which Greeks and Latins, spurious and genuine authorities sit side by side until they reach a stop marked “Renaissance”. Then grim-faced humanists climb aboard, check tickets, and expel fakes in hordes through doors and windows alike.’Footnote ⁵⁶ Filling the cultural vacuum which Roman pseudepigrapha has been made to inhabit means stepping out of the train-conductor role and treating these texts not as crime scenes but as creative responses to the literary past. Studying spurious Roman poetry as a literary tradition – not as an inextricable bundle of problems of authorial identities and dating – implies a fundamental repositioning of one’s scholarly stance towards marginal texts but also canonical authors, a willingness to drop the role of policeman of the canon and assume instead the mask of cultural observer.

Secondly, it is important to take stock not just of how approaches to the canon shape our approach to the margins, but of how the project of the canon is shaped by a given approach to the margins. The canonical locus of study for non-canonical texts has typically been the textual edition and the commentary and to this day pseudepigrapha and other ‘minor’ texts are mostly studied in the context of these most traditional philological genres.Footnote ⁵⁷ It is worth asking why this should be so and to what extent the scholarly medium, if at all, is shaping the message. From Scaliger’s edition of the Appendix Vergiliana (1572), to Baehrens’ Poetae Latini minores (1879–86), to Riese’s (1869–70) and Shackleton Bailey’s (1982) Anthologia Latina, to Housman’s and Goold’s editions of Manilius, engagement with minor Roman poetry has been surprisingly central in the tradition of textual criticism with genealogies of critics displaying their prowess and belonging to a scholarly tradition through successive and layered engagement with the inert bodies of marginal texts.Footnote ⁵⁸ As Lowell Edmunds argues, ‘minor literature’, which these texts are seen to embody, is understood squarely in aesthetic terms.Footnote ⁵⁹ Housman expressed this paradox lapidarily in a paper on the text of the Culex: ‘Just as it is hard to tell, in Statius or Valerius Flaccus, whether this or that absurd expression is due to miscopying or to the divine afflatus of the bard, so in the Culex and Ciris and Aetna it is for ever to be borne in mind that they are the work of poetasters. Many a time it is impossible to say for certain where the badness of the author ends and the badness of the scribe begins’ (Reference HousmanHousman 1902: 339). Here the textual work of the critic, often framed in clinical terms as a rescue operation or restorative surgery, constitutes an affirmation and sublimation of the method of philology, as the ‘aesthetic pleasure that comes from the sheer technical work’.Footnote ⁶⁰

It is also true, however, that the very method of traditional philology, which is being affirmed as the message, derives its prestige from the authority and value of the canon. The basic principle, common to both ancient and modern criticism, according to which the critic is able to judge on issues of authenticity because of his command of an author’s style (diuinatio), rests on the loving identification with the author.Footnote ⁶¹ The critic’s most important weapon is his acquaintance with each writer’s most salient characteristics and his ability to use his judgement and knowledge to spot the intruder. Modern authenticity criticism has made the stylistic variable arguably more ‘quantifiable’ by closely studying metrics and diction and compiling lists of features which are characteristic of a given author, and distinguishing between the early and late phases of their production. Yet, to ask whether a textual variant is authentic is to judge whether it reflects what are assumed to be the standards and style of the author, that is in effect to think like the author. Some conclusions follow: the authority of the critic is intimately dependent on the value of the reconstructed text. Unsurprisingly, minor texts pose challenges precisely in that they threaten to disrupt the critic’s desire to be one with the author. On the one hand, the minor text is an inert body, a kind of dummy on which the critic practises their surgical skill; on the other, as an inferior text, it must be resisted by the critic as an object of desire and identification.

However, there is one further corollary that we may be less eager to grapple with: namely that the project of the canon ultimately needs the presence of a non-canonical margin to justify itself. Taking a metaphor from the history of the book, we might think of the rejected readings in the apparatus at the bottom of a page as the paratext of philology. In Gérard Genette’s formulation, paratexts such as titles, dedications, prefaces, etc. represent the outer edge of the text and, though physically liminal, hermeneutically they are in fact central to helping configure and thus deliver the text.Footnote ⁶² This paradox of textual criticism has been well studied by critics of philology such as Jerome McGann and David Greetham. As the latter writes, in relation to the role of the apparatus, ‘the authentication of the “primary” text above is wholly dependent on the description and evaluation of the “rejected” readings below. The editors must successfully demonstrate the inadequacies of the lower text in order to convince the reader that the upper text is authentic.’Footnote ⁶³ In other words, the presence of the text above is guaranteed by, and therefore draws its authority from, the demonstrated inferiority of the lower text. The more inadequate the lower text, the higher the authority of the upper text. We can extend this metaphor beyond the sphere of textual criticism if we see this rhetoric of inadequacy and inferiority as fundamental to the self-definition of the canon. We may choose to preserve the status quo by seeking the admission of marginal texts into the canon (thereby implicitly marginalising other texts). Or we may choose to disrupt, if not the canon as a list, at least its claim to primacy and alterity versus the margin.

The Antiquity of the Canon

However much we may reposition ourselves in relation to both new and traditional canon formations, we must at some point grapple with the fact that the canon and the hierarchy of texts and genres on which it is built are to some extent inventions of Graeco-Roman antiquity. What does it mean to work with canon formations in the study of Latin literature given how the Romans themselves set out to construct a national literature as a translation of the canon of another?Footnote ⁶⁴ To what degree is the critical discourse of the Latin canon rooted in the very sense of belatedness and marginality built into the Roman literary tradition? To what extent does it feed upon and complement the classicism of the Second Sophistic under which the turn to the Greek classical past functioned both as a site of cultural resistance by the Greeks and as a tool of imperial control by the Roman elites?Footnote ⁶⁵ And to what extent does it depart from it?

There is a unique and intimate connection between Roman literature, which starts with a translation of a Greek ‘classic’, Livius Andronicus’ Odyssia, and the canon.Footnote ⁶⁶ This structuring of a Latin vernacular literature as a translation and continuation of Greek literature is predicated upon the existence of a more or less fixed canon of great works. In Denis Feeney’s words, ‘a strongly classicizing and canonizing urge is evident from the beginning of the translation project’.Footnote ⁶⁷ According to the standard narrative, the Romans of the second century bce inherited such a ‘canonising urge’ from their near contemporaries, the Hellenistic scholars and poets, who were the first to systematise the study of the Greek literary past. This narrative of translation and adaptation should be revised to account for the ways in which the Hellenistic systematising of the Greek literary past was interpreted through and made to map onto Roman modes of political and social ordering.Footnote ⁶⁸ Thus the Roman sources that discuss the work of Hellenistic grammarians can be read not just as storehouses of information about the third and second century bce but also about contemporary concerns with the appropriation of Greek heritage and the globalising aspirations of the Romans. In this way, the construction of the Greek classical past can be seen as an invention of the Romans as much as of Hellenistic Greeks, and one which created an enduring legacy in modern accounts of the literary canon.

The effort to frame Roman literature as a translation of the Greek canon is evident from the earliest beginning. Ennius was supposedly referred to as the alter Homerus already by Lucilius.Footnote ⁶⁹ The early Roman dramatists focused not on contemporary theatre but on the fifth-century ‘classics’.Footnote ⁷⁰ Cicero expresses this ‘canonising urge’ in a well-known passage of the Academica in which he is attempting to sell the project of writing philosophy in Latin:

Ennius, Pacuvius and Accius correspond to the tragic Greek triad of Aeschylus, Sophocles and Euripides, as already proto-canonised in Aristophanes’ Frogs. In turn, Plato, Aristotle and Theophrastus are presented as the philosophical canon.Footnote ⁷¹

The creation of a vernacular literature seems to have been accompanied by philological work that attempted to frame the developing body of texts as a match to and replica of the existing Greek classics.Footnote ⁷² Gellius, for example, transmits a fragment in iambic senarii from the De poetis of Volcacius Sedigitus, probably composed around 100 bce.

As Citroni notes, Caecilius, Plautus and Naevius constitute the initial triad, while the other six are added as an unwilling appendix (v. 12 Si erit quod quarto detur) to form a canon of nine.Footnote ⁷⁴ Ennius is amusingly presented as a tenth extra. Volcacius compares the critical process to an athletic competition: critics are fighting (v. 1 certare) and uncertain to whom they should award victory (v. 2 palmam deferant). This is not just yet another application of the ubiquitous metaphor of criticism as a form of contest. Rather, the competition with the conferral of the palm of victory is here emphatically seen as belonging to the Greek world, as Livy reminds us when discussing the introduction of the palma at the Ludi Romani.Footnote ⁷⁵ Varro’s own comic list was structured in triads depending on different areas of expertise with Caecilius being the best in plots, Terence in character and Plautus in dialogue (Sat. Men. 399 In quibus partibus in argumentis Caecilius poscit palmam, | in ethesin Terentius, in sermonibus Plautus).Footnote ⁷⁶ This canonising practice was later to be mocked by Horace in the Epistle to Augustus (Hor. Epist. 2.1.51 ut critici dicunt): as the critics say, Horace writes, Ennius is the Latin Homer, Afranius the Roman Menander, Plautus the Roman Epicharmus, while Caecilius is superior in seriousness (grauitas) and Terence in art (ars).Footnote ⁷⁷ Horace ironically mocks the rigidity of these expert judgements, while at the same time criticising the in-built bias against more recent writers. Paradoxically, however, Augustan literature can be read as an attempt not to do away with the archaic canon enshrined by Varro but to ‘reboot’ it, revamping with it the ‘canonising urge’ with which Roman literature had begun. When Horace at the end of Carm. 1.1 asks Maecenas to include him among the lyric poets (Quodsi me lyricis uatibus inseres), we are told that ‘Horace’s inseres represents the Greek ἐγκρίνειν, to “include in the canon” (οἱ ἐγκριθέντες, the classici)’.Footnote ⁷⁸ Here the Romans are seen to transfer pre-existing Greek cultural media that predated the advent of Rome – be they a list of canonical works, scholarly or poetic genres or institutional contexts for the preservation of poetry and learning. It is typically argued that this list of the nine lyric poets was an early example of canonical formation, created by the Alexandrian grammarians to identify the foremost writers in each genre.Footnote ⁷⁹ Yet, while an anonymous epigram from the Palatine Anthology seems to treat the canon of nine as closed (9.184 ‘beginning and end limit of all lyric’), another one plays with the conceit that Sappho is not the ninth lyric poet but instead the tenth muse (9.571).Footnote ⁸⁰ Neither is datable for sure to the Hellenistic period.Footnote ⁸¹ The first attestation of the canon of ten Attic orators is in the second-century treatise On Literary Styles by Hermogenes (2.401, 403) and forms the basis for pseudo-Plutarch Lives of the Ten Orators.Footnote ⁸² Most scholars agree that this list of orators (not works), just as the list of top tragic and comic writers, remained relatively fluid well into the late Republic when Cicero seems unaware of it (see, for example, Brut. 285–91).Footnote ⁸³

It is hard to date these Hellenistic lists, let alone ascertain their function and specifically whether they functioned as ‘proto-canons’:Footnote ⁸⁴ were Callimachus’ Pinakes (‘Tables’) a catalogue of books available in the library or a list of best in the ‘pragmatic sense of … most useful or most famous authors in the different genres’ (Reference EasterlingEasterling 2012: 286)?Footnote ⁸⁵ If the latter, it is unclear to what extent the Alexandrian enkrithentes are singled out as the best for the purpose of imitation, as their later counterparts discussed in Quintilian are, or simply as the most accomplished. The Latin word index, which translates Pinax, is used both of a catalogue of books from the library (Quint. Inst. 10.1.57 Nec sane quisquam est tam procul a cognitione eorum remotus ut non indicem certe ex bibliotheca sumptum transferre in libros suos possit) as well as a list of top surviving writers in a given genre or plays by a given author (Cic. Hort. fr. 48 indicem tragicorum; Sen. Ep. 39.2 indicem philosophorum; Gell. 3.3.1 the indices of Plautine plays by Volcacius Sedigitus). Either way, their circulation in written editions and commentaries of surviving Greek classical authors built towards and solidified their eventual canonical status.Footnote ⁸⁶

Above all, it is worth noting the extent to which this picture of the Greek canon as a pre-ordered hierarchical field is filtered through the Roman imperial imagination. To start with, not only is the list a quintessentially Roman cultural medium;Footnote ⁸⁷ Romans were also fond of displaying such lists of victories of comic or tragic poets in their homes as suggested by disparate findings of Athenian victory lists in Roman contexts.Footnote ⁸⁸ Regardless of when the lists of best writers such as those attributed to Volcacius Sedigitus and Varro were ultimately created, it is important to note the extent to which the account of the scholarly work of the Alexandrians is thoroughly mediated by Roman sources. Indeed, besides Horace, Quintilian is the most important source for the existence of these lists. According to the Roman rhetor:

Later in the reading list, Quintilian claims that ‘Apollonius does not appear in the grammarians’ list, because Aristarchus and Aristophanes, who evaluated the poets, included none of their own contemporaries’ (Apollonius in ordinem a grammaticis datum non uenit, quia Aristarchus atque Aristophanes, poetarum iudices, neminem sui temporis in numerum redegerunt, Inst. 10.1.54). It is worth stressing that these accounts of Alexandrian practice are deeply imbued with a typically Roman sensitivity to structure and social order.Footnote ⁸⁹ When, in the second century ce, Aulus Gellius (19.8.15) recalls Fronto’s way of calling the old cohort of orators or poets the classici, that is, men of the first class, he was creating a new coinage which was destined to have a long history, while all the same capitalising on a metaphorical interplay between literature and politics. As Farrell has remarked in relation to the patterning of poets’ careers onto the cursus honorum of Roman politicians, Hellenistic genre theory entered into a deeply productive dialogue with the political and social structures of the Roman upper class.Footnote ⁹⁰ Similarly, the Romans can be seen to have easily translated Hellenistic concerns with intellectual ordering into political hierarchies. It is no coincidence that in Quintilian’s description of the role of the Alexandrian grammarians the marking of inauthentic passages should be characterised as censoria uirgula notare (Inst. 1.4.3 ‘marking lines with a sign of disapproval’). This phrase is meant to bring to mind the nota of the censors, the mark that was put in the census next to the name of disgraced people to exclude them from the list of citizens qualified to vote and stand for office.Footnote ⁹¹ Donald Russell, whose translation I gave, renders the Latin ordo as ‘canon’ but the inclusion and exclusion of books from the list can be seen to mobilise the meaning of ordo as social class.Footnote ⁹² In Quintilian’s formulation, books are included or excluded from lists as citizens are included or excluded from classes (senatorial, equestrian, etc.) by grammarians who operate as Roman censors. Far from functioning as an inert metaphor, this intersection of the political and the canonical in the Roman literary imagination makes the canon an engine which both shapes and is shaped by Roman imperial ideology.

Another way in which the concept of the canon was already operative in Roman literature is in the clearly defined hierarchy of genres, adapted from the Greeks, according to which epic was seen as superior to all other genres.Footnote ⁹³ Consequently, several genres defined themselves as ‘minor’, ‘secondary’ or even para-literary. Epigram, for example, is at the bottom of the hierarchy.Footnote ⁹⁴ Prose is ranked lower than poetry with some exceptions.Footnote ⁹⁵ It is useful then to explore genre, canonicity and epigonality as mutually reinforcing cultural mechanisms within a carefully structured pecking order. Just as Martial’s defence of the low status of his epigrams (e.g. 1.praef. 7 iocorum nostrorum simplicitate; 11.20.10 Romana simplicitate) is well understood as a rhetorical strategy, not as a truthful characterisation of his style, so the epigonality of minor poets is being more and more understood as a self-conscious artistic effect.Footnote ⁹⁶ The anonymous author of the Laus Pisonis, addressing his would-be-patron, exclaims:

We can choose to take these statements at face value and let the author’s apologetic stance exercise its gravitational pull and claim, as one commentator does, that ‘the poet was only eighteen and whoever he was evidently had his way to make’.Footnote ⁹⁷ Or we can read these statements as part of a ‘generic pose’: either as a strategic rhetorical move designed to ingratiate the addressee or as a result of a conscious artistic persona not dissimilar from the ‘bumbling incompetence’ that Zetzel has seen as the hallmark of Horace’s poetic persona in the Satires.Footnote ⁹⁸ Lowell Edmunds’ invitation to isolate minor Roman texts not on the basis of aesthetic criteria but of shared literary effects such as parody and realism is very helpful. Could the minor and the marginal be approached as textual stratagems comparable to Roland Barthes’ ‘reality effect’?Footnote ⁹⁹ Conversely, following Mario Telò in his study of Aristophanic comedy, we may also choose to analyse the discourse of the canon in Graeco-Roman poets with a critical eye, not always assuming that it is ultimately dependent on an idealised list (e.g. a Hellenistic grammatical list). As Telò argues in relation to Aristophanes’ Peace, the canon of comedy is ‘an imagined, self-constructed ideal … deceptively offered as generic orthodoxy’ to undercut his rivals’ claims to poetic success.Footnote ¹⁰⁰

Latin literature offers a prime site from which to analyse the functioning of canons: for one thing, the Romans themselves saw their own literary corpus as a translation of the Greek canon. Though they inherited, translated and adapted the terminology of Hellenistic scholarship, they also uniquely shaped its intellectual tools: the modern discourse of the canon is not only rooted in the canonising project of Roman authors but also deeply connected from its inception with the ideological and political project of imperial conquest of the Romans.

The Boundaries of the Latin Canon

According to Cicero, Latin literature ‘began’ in 240 bce, when Livius Andronicus staged the first play at Rome (Cic. Brut. 71–3), but when and where does it end? The answer to this question involves analysing the complex relationship between the Early Church Fathers and classical antiquity against the rise of the discourse of secularism in modern academia and its impact on disciplinary fault lines. Accordingly, this section will problematise the structural distinction between classical and Christian Latin canon and emphasise instead the crucial role played by Christian sources, typically excluded from the study of Latin literature, in the shaping of the disciplinary tools, texts and reading practices of the average Latinist.

Since the birth of Altertumswissenschaft (‘the science of antiquity’) in the late eighteenth and early nineteenth century, the study of classical antiquity has defined itself in opposition to the study of coeval and co-extensive civilisations, chief above all the Egyptian and the Jewish.Footnote ¹⁰¹ While up to the middle of the nineteenth century, scholars moved with freedom between Graeco-Roman antiquity, early Christianity, New Testament and Hebrew Bible, these are today distinct disciplinary entities with their own autonomous canons.

These disciplinary distinctions inform current narratives of the rise of the canon. It has often been observed that while the word ‘canon’ is derived from the Greek kanōn, its modern usage with reference to a selective list of authors is a post-classical coinage based on a late Christian development, not unlike the ways in which the word ‘classic’ is, in Mario Citroni’s formulation, a ‘learned reuse, in the humanistic context, of a specific metaphorical expression used by an ancient author’.Footnote ¹⁰² In classical Greek, kanōn refers to a ‘rule’ or ‘standard’, and hence the term is used of authors who are presented as models for imitation. For example, we know from Pliny the Elder, Plutarch and Galen that the Greek sculptor Polykleitos assigned the name kanōn to a statue of perfect proportions and to a treatise on the subject of the ideal bodily ratio.Footnote ¹⁰³ In Dionysius of Halicarnassus, Lysias is said to be ‘the standard and model of excellence in this genre [i.e. narration]’Footnote ¹⁰⁴ while Thucydides is a ‘model historian and the standard of excellence in deliberative oratory’.Footnote ¹⁰⁵ Just as a canonical writer in the modern sense of the word is perceived to be superior, a kanōn is a model of recognisable value. Yet, in so far as they are held up above the rest, the kanones are models for the practical purposes of imitation. It is in the context of which works should be read and imitated by would-be-orators that the practice of listing model authors first begins to emerge.Footnote ¹⁰⁶ The pragmatic context of emulation creates what at first seems to be a critical difference between the ancient and the modern construct of a list of authors. Whereas, as we have seen, the canon in the modern sense of a group of preferred works hints at the trans-historical superiority of the selected works of literature, the ancient list is designed to point out the virtues of each model for the purpose of imitation in the context of the training of orators (Quint. Inst. 10.1.45 Sed nunc genera ipsa lectionum, quae praecipue conuenire intendentibus ut oratores fiant existimem, persequor). Such models are not, however, to be viewed a-critically; rather, the student is invited to improve upon the model, careful to reproduce its virtues but not its faults.Footnote ¹⁰⁷ In addition to accommodating a certain amount of fluctuation, these lists of proto-kanones are not list of works but of writers.

The word kanōn begins to be used of a closed list of texts by the early Christians in discussions of the authenticity of specific books. Thus in the early fourth century ce, Eusebius of Caesarea writes of Origen that ‘in his first book on Matthew’s Gospel, maintaining the Canon of the Church (τὸν ἐκκλησιαστικὸν φυλάττων κανόνα), he testifies that he knows only four Gospels [Matthew, Mark, Luke, John], writing as follows … ’ (Euseb. Hist. eccl. 6.25.3).Footnote ¹⁰⁸ In Eusebius, although there is continuity in the deployment of kanōn to isolate texts deemed to be endowed with specific virtues, the emphasis has clearly shifted from stylistic and ethical emulation to the process of sanctioning and listing works deemed morally and theologically superior. This meaning is well defined by the fourth century when Augustine and Jerome use canonicus, -a, -um of books belonging to a sanctioned body of work, in other words a canon. Augustine explains the origin of the canon thus understood as the Word of God in The City of God:

It is standard to draw a fault line between this Christian usage of the canon as a body of authoritative works and classical lists of kanones as well as the later post-humanistic repurposing of the term canon to secular writings.Footnote ¹¹⁰ However, it is important to note that these seemingly different canonical formations stand on a well-identifiable continuum: albeit rooted in a different kind of claim about the intrinsic superiority of the work, the modern usage of the word canon retains an emphasis on emulative practices through its appeal to the place of the canonical texts in the ethical formation of readers.Footnote ¹¹¹ Conversely, because of the close connection between style and character (qualis homo talis oratio: Cic. Tusc. 5.47.9; Sen. Ep. 114), the classical kanones have an unspoken potential to be used as ethical models from the very beginning.Footnote ¹¹² Thus despite the fact that it is often claimed that Christianity has superimposed its own canon-forming gaze onto antiquity, aesthetics and ethics are deeply interwoven in formations of canons from the very beginning.

This complex entanglement of church and text raises a related point about the disciplinary divide at the heart of this narrative. To glance back at the history of what we now call ‘classical scholarship’, Isaac Casaubon, the renowned sixteenth-century scholar, became famous for disproving the authenticity of the Corpus hermeticum, which on the basis of linguistic borrowings from the Septuagint, Plato and the New Testament, he proved to be not the product of its purported date of composition but of the first centuries of the common era.Footnote ¹¹³ To jump to the nineteenth century, the rise of textual criticism is deeply connected to parallel developments in the study of ‘sacred’ texts: indeed Karl Lachmann, from whom, rightly or wrongly, the stemmatic method takes its name, edited both the New Testament and ‘classical’ texts such as Lucretius, as did Richard Bentley before him.Footnote ¹¹⁴

How far back should we trace this disciplinary divide between classical and Christian Latin? The split between pagan and Judaeo-Christian literatures is arguably already encapsulated in Jerome’s often-cited dream, where God appears to him and accuses him of being ‘a Ciceronian, not a Christian’ (Ciceronianus es, non Christianus, Jer. Ep. 22.30).Footnote ¹¹⁵ Echoing Tertullian’s memorable question – ‘what does Athens have to do with Jerusalem?’ (quid ergo Athenis et Hierosolymis?, De praescr. haeret. 7.1) – Jerome prefaces his narrative of his own renunciation of classics by positing a radical split between classical literature (Horace, Virgil and Cicero) and Christian writings:

Of course this uncompromising stance in Jerome’s letter, written in 384, is not only belied by his literary practice, in which quotations of and allusions to classical literature abound, but also by later pronouncements. For example his letter 77, written in 398, in response to criticism from the orator Magnus who had accused Jerome of defiling the purity of the church with the filth of the pagans by his frequent quotations of secular literature (Ep. 77.2 cur in opusculis nostris saecularium litterarum interdum ponamus exempla, et candorem Ecclesiae, Ethnicorum sordibus polluamus), Jerome compares secular wisdom (sapientia saecularis) to the beautiful captive woman of Deuteronomy 21.35 for whom the law prescribes a ritual of cleansing and mourning before she becomes the wife of her captor. Jerome’s desire for the captive secular wisdom is justifiable in that it proliferates the faith and does so by cutting or shaving off whatever pleasure, error and idolatry is hiding in the classical texts.Footnote ¹¹⁶

Almost at the same time, Augustine lays out his programme for Christian learning in De doctrina Christiana, written in 396 to address the issue of the teaching of Christianity. In this important manifesto, Augustine operates under a strict dichotomy between ‘Roman’ and sacred authors: the latter are called canonical authors (canonici auctores) and encompass both sacred texts and church fathers. Much of the aim of book 4, devoted to the issue of Christian eloquence, is to show that facility in speech can be acquired by reading and studying the Christian canon (ecclesiasticae litterae):

While making a case for the stylistic value of Christian literature, Augustine happily quotes and follows Cicero’s theory of the three styles and his admonition that the orator be made to master all styles. After giving numerous positive examples of the different styles from both scripture and church fathers, Augustine concludes by calling for an acknowledgement of the eloquence of ‘our canonical authors and teachers’.Footnote ¹¹⁷ This radical distinction between secular and ecclesiastical was codified in the sixth century in Cassiodorus’ Institutiones diuinarum et saecularium litterarum, a treatise on Christian education in two books, the first on scripture and the works of Christian fathers and the second one on the secular liberal arts. Originally designed for use in the Vivarium, a monastic community he founded in his own family estate in Calabria, the book, variously excerpted and copied, was to be foundational in the Middle Ages.Footnote ¹¹⁸

The church fathers’ radical and self-conscious split between secular antiquity and Christian writing anticipates the disciplinary divide between Classics and theology but only to an extent. The modern secularisation of education and the loss of biblical knowledge have catalysed a contraction of the canon of Latin works taught and researched in Classics departments to works of the pagan era which, though more distant in time, are paradoxically more ‘legible’ to the modern secular reader, in so far as they are not imbued in a religious ideology that many do not consider relevant. To the extent that late antique literature has a place in the classicist canon, it tends to be approached in relation to its translation and adaptation of pagan antiquity. Despite its importance, this work privileges Christian learned poets like Prudentius, Ausonius and Claudian, while Patristics scholars focus on the study of the contemporary writings by the Early Church Fathers.Footnote ¹¹⁹ To some extent, this disciplinary divide inscribes a modern distinction between secular and religious onto antiquity, by relegating the study of the supposedly secular pursuit of ‘literature’ to the late antique scholars trained in the classical tradition, while reserving the early Christian theology for the theologians.Footnote ¹²⁰ There are obvious pitfalls to this modus operandi: on the one hand, for some time now, scholars have focused our attention on the pivotal role played by secular fields such as, for example, grammar, in the development of Christian textuality.Footnote ¹²¹ In turn, as we shall see in a moment, this work is a reminder that the creation of the classical past is unthinkable outside of the mediation of ‘Christian’ sources. In short, in approaching late antiquity we should be wary of modes of interpretation which, in Debora Shuger’s words from her study of the Renaissance bible, frame religion as a ‘separable “layer” atop the surface of [secular] culture’; rather it is best to approach all forms of knowledge in this period as the result of complex syntheses of religious and secular discourses.Footnote ¹²²

Moreover, the modern marginalisation of the late antique Latin corpus affects not just Christian texts that do not fit the definition of secular literature but also contemporary works written by pagan authors or by authors whose religious stance is difficult to ascertain, as for example is the case for Martianus Capella and Macrobius. Such a dichotomy either inhibits the study of tralaticious corpora (e.g. commentaries) or else generates scholarship aimed at disentangling historical kernels rather than situating the work as a whole.Footnote ¹²³ An interesting and complex example of this phenomenon is the Anthologia Latina, a compilation of Latin poems (elegies, epigrams, inscriptions, centos and other genres) assembled starting in the Renaissance with Scaliger on the basis of early mediaeval anthologies – principally the Codex Salmasianus (Paris lat. 10318, eighth–ninth century) and the Codex Thuaneus (Paris lat. 8071, ninth century).Footnote ¹²⁴ The collection in the Codex Salmasianus, which furnishes the bulk of the modern Anthologia Latina edited by Riese and others, originated in Vandal North Africa of the sixth century but it contains earlier material, as, for example, three epigrams attributed to the younger Seneca. The Salmasianus and the Thuaneus preserve three epigrams under the name of Seneca (Anth. Lat. 224, 228, 229 Shackleton Bailey), two of which are also found without any ascription in the Codex Vossianus, where they are followed by another sixty-nine poems. No name is attached to these epigrams, but on account of the similarity in style and content, editors have created a collection of (pseudo-) Senecan exile poetry.Footnote ¹²⁵ Yet the Codex Salmasianus can also be approached in its own right as a witness to Virgil’s reception in the scholastic environment of late antique Vandal Carthage in which earlier specimens of Virgilian pastiches such as the Medea of Hosidius Geta (end of the second/early third century ce) or the Pervigilium Veneris (of disputed date) co-exist with more recent ones.Footnote ¹²⁶

It may therefore be argued that the disciplinary divide between Classics and theology leaves much to be desired in terms of our ability to comprehend both the intellectual milieu from which the literary works of late antiquity – be their authors pagan or Christian – originated, and by extension the very corpus of texts which we comfortably group under the umbrella of the ‘classical tradition’. To give one example, the Saturnalia of Macrobius is now thought of as the product, not of the late fourth century, as previously assumed, but of the early fifth – Cameron identified the author of this work with Macrobius Ambrosius Theodosius, praetorian prefect in 431 ce.Footnote ¹²⁷ In Cameron’s dating of the Saturnalia to 431, after the establishment of Christianity as the religion of the empire and twenty or so years after the sack of Rome in 410, it is almost impossible to maintain that Macrobius was not a Christian. Yet, in so far as the work makes no mention of Christianity and is set some fifty years prior to its date of composition during the Saturnalian feast of 382 or 383, on the eve of Gratian’s abolishment of pagan cults in 382, it has been standard to approach the dialogue as a source for Virgilian scholarship of the earlier first and second century. Whose hermeneutics does the discussion of Virgilian poetry, which occupies books 3 to 6, reflect? That of the pagan elites of the characters in the dialogue, such as Quintus Aurelius Symmachus, one of the dialogue’s main characters, a successful orator, statesman and leader of the unsuccessful protest against Gratian’s removal of the Altar of Victory? Or that of the early imperial sources? Or of Macrobius’ own Christian readers? Take, for example, Macrobius’ much contested approach to Virgil as a sacred shrine, in which Praetextatus lays the ground for the exegetical discussion of the poet that follows:Footnote ¹²⁸

The image of commentary here deployed is spatial: interpretation is compared to an entrance into the inner space of a sanctuary (adyta) and hermeneutics to opening up a sacred shrine. It may be that in Cameron’s words we should resist the notion that Macrobius ‘saw Vergil as a sacred text, from which he planned to extract arcane religious truths’.Footnote ¹²⁹ One could easily trace this language earlier on, to, for example, Tacitus’ portrayal of the inner shrines of eloquentia.Footnote ¹³⁰ A similar question arises about Virgil’s divine status. The earlier attestation of the title diuinus for Virgil is in Catalepton 15, a book of pseudo-Virgilian iuuenilia generally dated to the first century ce. In presenting the preceding collection, the poem acts like a sphragis of sorts, stating that ‘these too are the first beginnings of that divine poet’ (Catal. 15.3 illius haec quoque sunt diuini elementa poetae). Yet when this title is repurposed in the Servian commentaries, how are we to interpret the statement that ‘the divine poet always touches on the truth, even when he is engaged in some other matter’ (unde apparet diuinum poetam aliud agentem uerum semper attingere, Serv. Dan. 3.349)? Is the commentary’s approach to Virgil as ‘all full of knowledge’ (totus quidem Vergilius scientia plenus est, Serv. ad Aen. 6 praef.) harkening back to classical antiquity or forward to Christian allegoresis?Footnote ¹³¹ Servius, whose commentary on Virgil probably predates the sack of Rome of 410, and who is also a character in Macrobius’ Saturnalia, makes liberal use of allegory, a tool developed in classical antiquity but also associated with religious exegesis, both Christian and Jewish.Footnote ¹³² Yet it is surely confining to read ‘sacred’ as reflecting exclusively a supposedly hidden Christian religious agenda.Footnote ¹³³ Rather, the question to ask is how Macrobius’ or Servius’ repurposing of the earlier language of the sacrality of poetry might resonate with his readership, Christian and otherwise – how in effect the religious experience of reading communities both leverages earlier representation of Virgilian and Homeric omniscience and reshapes them.

Finally, the chief disciplinary tools we use in our fields and those without which the classical tradition is materially unthinkable – the commentary, the textual edition and translation – were developed in the context of the exegesis of scripture, which was in itself read in dialogue with classical texts.Footnote ¹³⁴ Many of our Graeco-Roman sources are known exclusively or almost exclusively through indirect tradition and transmitted in much later sources which have their own ethics of quotation.Footnote ¹³⁵ The Latin canon presents distinctive challenges in this respect: whereas papyri, mostly from Egypt, have yielded a vast quantity of fragments from Greek literature, the vast majority of fragments of Latin literature survive as quotations.Footnote ¹³⁶ Thus late antique grammatical and exegetical sources played a pivotal role in the transmission of republican literature.Footnote ¹³⁷ In other cases, the mediation of Christian writers was instrumental in the survival and inseparable from the transmission of a given work: in the case of Cicero’s Topica, for example, the text survives largely accompanied by a commentary on the text by Boethius written around 500 ce. The Boethian commentary rivalled in popularity the Ciceronian original, the tradition of which it contaminated throughout the Middle Ages.Footnote ¹³⁸ In a sense, therefore, the distinctions between scriptural and profane, classical and late, though in principle correct, threaten to foreclose a deeper conversation about the shaping of texts in the context of tradition, reading and interpretation.

The Latin Canon: a Brief History

The canon of Latin literature is far from a straightforward entity. As we are about to see, its two constitutive elements – Latin and literature – are both subject to negotiation. On the most straightforward interpretation, a Latin canon would include only texts written in the Latin language. Yet, ‘Latin’ is a highly problematic term and susceptible to a variety of readings, depending upon which highly different canonical formations might be construed or imagined. The traditional and most conservative version of the canon of Latin literature begins with Livius Andronicus, Naevius and Ennius and ends with Tacitus and Juvenal and the authors of the ‘High Empire’, namely Apuleius, Aulus Gellius, Fronto and the poetae nouelli of the Antonine Age.Footnote ¹³⁹ This structure is already present in Friedrich Leo’s pioneering Geschichte der römischen Literatur, published in Berlin in 1913, and by many considered the founding text of Roman literary studies.Footnote ¹⁴⁰ Although Leo died after the publication of the first instalment (‘Die archaische Literatur’) and the project was never completed, one can get a glimpse of its ambition from the treatment of Latin literature in Wilamowitz-Moellendorff’s Die griechische und lateinische Literatur und Sprache, published in Hinneberg’s Kultur der Gegenwart, a massive multi-volume history of world culture which came out between 1905 and 1926. There ‘Die lateinische Literatur und Sprache’ (‘Latin literature and language’) is divided into three sections: one, written by Franz Skutsch, on the history of the Latin language; one, authored by Leo, on Latin literature of antiquity; and one, written by Eduard Norden, on Latin literature in the transition from antiquity to the Middle Ages. Norden’s treatment of the transitional period, which partially overlaps with Leo’s, examines the development of literature from the mid-fourth century, the time of ecclesia triumphans, to the reconstitution of the Roman empire under Charlemagne but is careful to acknowledge that this history is partial and does not address church history (p. 484). Leo’s treatment of Latin literature is chronologically arranged in three main periods: republican, Augustan and imperial, though the first section is almost equal in length to the sum of the last two. The rationale for this choice is made explicit in the beginning of the section on imperial literature: after Augustus, the Romans did not produce any literature of equal value; among the post-Augustan poets, there are some great names, also some with considerable talent but no great poets (‘sind große Namen, auch beträchtliche Talente, aber keine großen Dichter’, p. 454). While this history of Latin literature follows the seemingly self-evident contours of chronology, Leo’s emphasis on Rome as the spiritual link (‘das geistige Band der alten und neuen Weltkultur’, Reference LeoLeo 1913: 1) between the old (Greece) and the new (Christian Europe) tailors a specific approach to Latin canon, one that privileges acts of transference and adaptations (hence Leo’s intense focus on the earlier phases of Roman literature). The decadence of Roman literature is explained as the by-product of the rise of rhetoric and the metamorphosis of poetry into rhetorical art (‘rhetorische Kunst’). Seneca, Juvenal, Martial and Tacitus are the highlights of this age, while Lucan and Statius are among the considerably talented but their work is dismissed within a few words. In regards to Statius, the Thebaid is praised as lacking ‘neither power nor impetus’ (p. 459) but neither the Siluae nor the Achilleid find any mention. The literature of the later empire includes the archaising movement of the Hadrianic era which with its backward-looking glance is said to have led to the collapse of Roman literature and the rise of the provinces and the lower classes, with Africa being singled out as the cradle of the novel (Apuleius) and of the first translations of the Bible (Tertullian, Augustine, etc.). There is a brief mention of Christian poetry and the section on antiquity ends paradigmatically with a treatment of Boethius’ De consolatione philosophiae, a prison dialogue replete with classical learning and written as its author, born shortly after the deposition of the last Roman emperor, awaited execution by the Ostrogothic king Theoderic. Later to become one of the most popular works of the Middle Ages, in Leo’s survey of Latin literature the De consolatione philosophiae comes to epitomise the end as well as the enduring legacy of Roman antiquity.

In order to understand Leo’s selective focus and chronological limiting of Roman literature to the fall of the Roman empire, it is helpful to glance at earlier stages in the development of the study of antiquity. Leo’s approach to the question of the originality of Latin letters was in many ways a response to German philhellenism which, starting with Winckelmann in the late eighteenth century, had elevated the Greeks and disparaged the Romans as mere imitators. And yet philhellenism in its complex relation to nationalism is in many ways responsible for the birth of Latin literary history. For it was Friedrich August Wolf, the father of Altertumswissenschaft, who authored the first history of Roman literature in parallel with the founding of the seminarium philologicum in Halle in 1787.Footnote ¹⁴¹ Wolf’s ‘History of Roman literature’ is arranged in two parts. The first is a survey of known authors arranged under the names of rulers and emperors. The second is a history of the genres of Roman literature: poetry (dramatic, epic and lyric) and prose (historiography, eloquence, erudition), which surveys relevant authors in chronological order.Footnote ¹⁴² This dazzling work, which still repays reading, begins with a theoretical discussion of literary history defined as ‘the coherent narrative of the fortunes of the scientific and learned enlightenment of a people’ (‘eine zusammenhängende Erzählung von den Schicksalen der wissenschaftlichen und gelehrten Aufklärung einer Nation’, Reference Wolf and GürtlerWolf 1834: 3). This literary history is not a collection of facts but a narrative with a story to tell about the rise and fall of a nation. The canon therein constructed is thus subservient both to this narrative as a whole and to the operative concept of people: hence, for Wolf the chronological limit of Roman literary history is the deposition of Romulus Augustulus in 476 (p. 94), though some of the authors mentioned in the second part are later (e.g. Isidore and Beda). Authors writing between the demise of Latin knowledge in the sixth and its renaissance in the fifteenth century are listed in a separate appendix. Secondly, a clear connection is made between political and literary decline: the period that goes from Hadrian to the sack of Rome witnesses at once the crumbling of state institutions and the demise of literature.Footnote ¹⁴³

In order to appreciate the ideological thrust of Wolf’s narrative, it is useful to compare it to that of his model (see p. 8), the Bibliotheca Latina of the seventeenth-century polymath Johann Fabricius published in Hamburg in 1697, the most comprehensive bibliographical work of its generation which lists all known authors of the classical period as well as editions and translations, and gives a critical discussion of each work. This original edition of the Bibliotheca begins with Plautus and Cato the Elder, includes late antique writers like Macrobius, Claudian and Symmachus, and ends with Boethius and his younger contemporary and successor Cassiodorus. An appendix follows containing a list of fragmentary poets (beginning with Ennius), a list of Christian poets and one of ecclesiastical writers – his principal stated aim being that of providing an overview in chronological order of the totality of Latin authors, the ancient ones but not the ecclesiastics (latinos autores veteres non ecclesiasticos, quorum scripta aetatem tulerunt, recenserem universos ordine chronologico, Reference FabriciusFabricius 1697: 3). The rationale given is primarily pragmatic rather than ideological – there are many editions of ecclesiastic writers available to those who are interested in this area of study such as for example Cardinal Bellarmino, De scriptoribus ecclesiasticis (1613).Footnote ¹⁴⁴

Even if we chronologically restrict the focus of study, the canon of Roman literature includes on most definitions at least some works not written in Latin: the earliest Roman historians, the annalists Fabius Pictor and Cincius Alimentus in the second century bce, wrote in Greek. Historians of Rome also contend with the fact that Roman history written by provincial Roman elite continues to be written in Greek: to name just two notable examples, in the early second century ce, Plutarch of Chaeronea, a Roman citizen, wrote in Greek the Parallel Lives, a comparison of Greek and Roman history, as did Cassius Dio, a Roman senator from Bithynia who wrote a work of Roman history in Greek in the early third century ce.Footnote ¹⁴⁵ The slippage here is due to the fact that ‘Roman’ denotes a political identity, shared by speakers and writers of several languages including, but not limited to, Greek and Latin (see Goldhill, Chapter 16 in this volume).Footnote ¹⁴⁶ Latin, however, need not imply Roman: in fact, Ennius, a speaker of Greek, Latin and Oscan, only received Roman citizenship through Marcus Fulvius Nobilior in 184 bce; Terence, whose cognomen was Afer (‘the African’), was allegedly born in Carthage (Suet. Vita Ter. 1), enslaved and brought to Rome to the home of the senator Terentius Lucanus in the 160s bce. Thus one might question to what extent other literature written in Greek from later periods might or should also have a place in a history of Roman literature: a notable case study in this category might be the epigrams written by Greek poets for Roman patrons, like Antipater of Thessalonica’s poems for and about Lucius Calpurnius Piso, which would have been collected in the Garland of Philip, an anthology of first-century-bce epigrams compiled some time before the death of Nero and now transmitted in the Greek Anthology, and other writings in Greek by literati belonging to the same circle (e.g. the works of Philodemus, the remains of which were found in a villa in Herculaneum which supposedly belonged to Lucius Calpurnius Piso’s father, the consul of 58 bce, and Parthenius of Nicaea, whose Erotika pathemata is dedicated to Cornelius Gallus and may have been influential with the neoterics and elegists).Footnote ¹⁴⁷ Other influential works in the study of Roman literature and history, such as the Res gestae, exist in multiple parallel versions (Greek and Latin) recovered from the edges of the empire (Ancyra, Antioch and Apollonia).

As scholars become more interested in and aware of the hybridity of literary traditions in the imperial period, this sub-canon of Greek/Roman/Latin works is bound to increase in size and challenge long-standing assumptions about literary interactions in the period.Footnote ¹⁴⁸ To push the issue further, if one were to renounce, or at least resist, the notion of ‘Greek’ and ‘Roman’ as self-contained concepts, one might construct far different canons, ones based, for example, on genres or provincial and local identities. Hence studies of the novel have already created hybrid canons, consisting of works in Latin such as Petronius’ Satyricon or Apuleius’ Metamorphoses, which presents itself as derived from a Greek original (1.1 fabulam Graecanicam), in Greek (Chariton, Achilles Tatius, Longus and Heliodorus) as well as texts that appear in multiple redactions (e.g. the Alexander Romance, which survives in Greek and Latin, as well as Syriac and Arabic) or Latin texts that purport to be translations of Greek originals (e.g. Ephemeris belli Troiani of Dictys of Crete or the Historia Apollonii regis).Footnote ¹⁴⁹ Local identity, whether that is understood to refer to authors or readers, might also be deployed as an organising principle to generate alternative canonical formations.Footnote ¹⁵⁰ Scholars have long remarked upon the Spanish school of the Senecas, Lucan, Martial and Quintilian or that of North Africa exemplified by Apuleius, Fronto and the Christian apologetics of the second to third centuries like Tertullian, Cyprian and Minucius Felix, as well as Claudian and most notably AugustineFootnote ¹⁵¹ – though less obvious might be the ways in which the local affiliations of authors might interplay with displaced audiences of provincial readers in the various peripheries of the empire.Footnote ¹⁵²

Moreover, one’s definition of ‘literature’ inevitably yields a different picture of the canon. The canon of eighteenth- and nineteenth-century literary history is structurally centred on the high genres of poetry, the apex of which marks the high period of Roman literature. By contrast a history of Latin literature centred on a canon of Latin legal writings would begin with the law of the Twelve Tables and the fragmentary remains of its early commentators, such as Sextus Aelius in the second century bce, and extend all the way through the high classical period of Roman law in the second century ce with the works of Salvius Iulianus, commissioned in 131 ce by the emperor Hadrian to edit the Praetor’s edict, Ulpian, Gaius, the author of the Institutiones, and Sextus Pomponius, stretching all the way into the sixth century and culminating with the Corpus iuris ciuilis, a magnificent compilatory work of known Roman law commissioned by the emperor Justinian in 527 ce and a source for many earlier legal sources. Conversely, a canon of Latin grammatical writings would yield a different picture altogether, beginning as Zetzel’s guide does, in the second century bce with exegetical work on Ennius and Plautus by Lampadio, Aelius Stilo, Volcacius Sedigitus and others, moving to Varro in the mid-first century bce and to the archaist movement, and culminating with the ‘classical era of Roman philology’ – the centuries between the fourth and the sixth bookended by the commentaries and grammatical works of Marius Victorinus, Aelius Donatus and Servius (end of fourth–early fifth century ce) to Priscian, Boethius, Cassiodorus in the sixth century and Isidore in the early seventh century, whose massive Etymologiae is both a compendium of earlier pagan grammatical work and often considered a watershed between late antique and mediaeval grammar.Footnote ¹⁵³ As social and cultural history become accepted components of literary study, not only the boundaries of the canon are likely to shift but also the narratives of its development. As we have just observed in the cases of the legal and grammatical corpora, historical moments considered to mark peaks and declines in the narrative of the high literary canon – the classicism of the age of Augustus, the silver age of the Neronian and Flavian age, etc. – do not appear to constitute significant shifts in these other canonical formations. Additionally, one must also guard against the tendency to construct chronological narratives of literary development loosely based on the succession of genres. Thus few would agree today with the view of post-Ovidian literature as the ‘age of rhetoric’ or with late antiquity as the ‘age of technical knowledge’, labels that prioritise poetry and some genres of prose (historiography, Ciceronian oratory, etc.), while all the while eliding from view both the long history of rhetoric in Rome beginning in the second century bce and the development of poetic genres in late antiquity (to name just two phenomena).Footnote ¹⁵⁴

It is important to note that while the ancient hierarchy of genres certainly informs the discourse of canon, it is not coincidental with it. Thus authors whose works occupy a low place in the hierarchy of genres have at times occupied an important place in the canon. Statius and Lucan were immensely popular in the Middle Ages with hundreds of surviving copies but were marginal at best throughout the twentieth century, a fact memorably acknowledged by Statius’ Oxford editor when he stated that ‘there is such an abundance of manuscripts of the Thebaid that one rightly suspects that Statius had more copyists during the Middle Ages than readers in our time’.Footnote ¹⁵⁵ The hazards of survival are sometimes responsible for changes to the canon: Catullus was lost in the Middle Ages and only came to light around 1300 in Verona. Lucretius was largely unknown until Poggio Bracciolini’s discovery of a manuscript in 1417 during the Council of Constance, from which point it arguably shaped not just the Renaissance’s reception of antiquity but the very concept of modernity. A re-dating of an author can also engineer a turn in the fortune of their text: rediscovered by Petrarch in 1360, the Eclogues of Calpurnius Siculus were considered contemporary to the poems of Nemesianus, whose work is transmitted in the same manuscript and who was then identified with the third-century-ce author of the Cynegetica. Since 1854, when Moritz Haupt redated the Calpurnian eclogues to the Neronian period, Calpurnius has frequently appeared in literary surveys of the Neronian age, despite the challenges mounted to Haupt’s arguments by Champlin and others.Footnote ¹⁵⁶ In turn, while in contemporary discourse, the canon refers to the list of works studied and perhaps researched in schools and universities, at a time when Latin was widely spoken and read, the school canon was not necessarily coincidental with that comprised of works privately read, let alone translated.Footnote ¹⁵⁷ Some texts barely known to specialists today were immensely popular and have complex manuscript traditions that defy neat divisions between antiquity and the Middle Ages: these include epitomators of the Roman historians such as Justinus, the epitomator of Pompeius Trogus, ‘a household name’ through the Middle Ages, and Eutropius, whose Breviarium was reworked by Paulus Diaconus in the eighth century and by Landolfus Sagax in the tenth.Footnote ¹⁵⁸ In this class also belong the Distichs of Cato, a moralising work from late antiquity attributed to Cato the Elder, which was a wildly popular school text and translated already into Old English as early as the twelfth century. Finally, one must not forget that literature is in itself a relatively recent invention. To give just one example, in Oxford from its establishment in 1800 to 1972, the course of study known as Literae Humaniores – the ancestor of today’s Classics degrees – was largely dominated by history and philosophy with Latin and Greek being tested through prose composition and unprepared translation.Footnote ¹⁵⁹ If anecdotes are at all effective at conveying the lived reality of the canon, when glancing at nineteenth-century autobiographies one is struck by the frequent mentions of history and antiquarianism at the expense of what we may today identify as canonical authors. Henry Fynes Clinton, a member of Parliament and a classical scholar who studied in Oxford between 1799 and 1806, left a rich autobiography. In it he writes that when he was in school, he read ‘Virgil, except the Georgics; almost the whole of Horace, the Gallic War of Caesar; Sallust, and the Catilinarian orations of Cicero’.Footnote ¹⁶⁰ By the time he left Oxford, ‘except the orations, read at Southwell [school], I had not studied any parts of the works of Cicero. I was ignorant of Quintilian, and Tacitus, and Pliny. I had twice perused Livy with attention.’Footnote ¹⁶¹

Conclusion

Moving between a focus on the physical limits of the canon across time and a review of the canon’s impact as a critical mechanism, we have discussed the ways in which the very discourse of the canon has traditionally influenced scholarly approaches in the field of Latin literature and Classics broadly understood. Emphasis has been placed on the ways in which the canonised nature of the classics determines not just a hierarchy of texts and methodologies worthier of being taught and researched but also informs the very approach to non-canonical or ‘para-canonical’ texts. The canon in other words is not just about what we study, it is also about how we study it. Interrogating the canon of Latin texts implies a fundamental repositioning of one’s scholarly stance not just towards non-canonical texts but also towards canonical authors, whose primacy should be scrutinised.

No one will dispute that the Latin canon, both in its internal structure and as a mechanism for organising knowledge, has come under intense scrutiny. But what kind of philology can we imagine outside of the discourse of the canon? And how, if at all, will the study and theoretical framing of so-called ‘marginal texts’ change the discipline?Footnote ¹⁶² It is critical that we modify not only our academic cores and scholarly foci to include a wider complement of texts, ancient and post-classical, but also that we change the frame of reference within which marginal texts are typically considered. This shift will involve in the first instance an awareness that the marginal status of such texts cannot be elided from any story we tell about them. We cannot apprehend marginal texts in spite of themselves and outside of the rhetoric both ancient and modern that has constituted them and at times relegated them to the periphery. Protestations of poetic incompetence, inferiority and epigonality are a trope in ‘minor’ texts but one that we would be well advised to resist. Statius’ famous wish expressed at the end of the Thebaid that his poem ‘not challenge (ne tempta) the divine Aeneid but rather follow it from a distance (longe sequere) and ever worship its footsteps’ (Theb. 12.816–17) embodies both the canonical hold of the Virgilian epic and the self-proclaimed inferiority of his successors.Footnote ¹⁶³ Yet, as Philip Hardie notes, it also ‘sanctifies’ the canon into which the poet seeks admission in the near future: ‘even if Statius advises his poem for the present to follow at a distance, the future holds honores that might well be those of a god’.Footnote ¹⁶⁴ Moreover, scholars of post-Virgilian epic have pointed to the ways in which the self-conscious epigonality of texts and authors functions alternatively as a cover for critiques of their predecessor (Reference GanibanGaniban 2007) or as a productive mechanism for exploring a new balance between periphery and centre in the changed political environment of the Flavian era (Reference AugoustakisAugoustakis 2010).

We may choose to study this discourse of canon and margin diachronically and effectively historicise the canon, not simply tracking items that have dropped in and out of it, but above all examining its changing function in history, its connection to forms of institutionalised knowledge (the Hellenistic library, the German university, etc.) and how canonical formations develop side by side with other social and intellectual structures. Or perhaps marginal texts will lead to a renewal of the discipline if we can frame marginality not as a list of physical objects (minor texts) or as a physical ‘no go space’ that needs to be rescued into the centre but as a modus operandi. In proposing marginality as a ‘way of doing business’, I have been inspired by the possibility of looking at the margin (or marginalisation) as a process rather than as object. One of the most acute contemporary readers of the western margin, the subaltern and the other, feminist and cultural critic Gayatri Spivak, once wrote that instead of pointing the accusing finger at the centre, we would do better to use ourselves ‘as a shuttle between the centre (inside) and the margin (outside) and thus narrate a displacement’ (Reference SpivakSpivak 1987: 146). A new ethics of the periphery, as one might call it, would leverage the liminality of non-canonical texts to interrogate the ideological pressures that shape practices of reading in and out of the canon and to engage with and document processes that led to the marginalisation or exclusion of different texts at different moments by different readers. It is this ability to conceive of bold ‘narratives of displacement’ between canon and margin rather than reproduce new descriptions of the margins according to old narrative patterns that will ultimately open up a new way to imagine the literary space of antiquity.