I

Up until the 1960s, most studies of Montaigne had tended to invoke the Essais as an autobiographical source, a window into the essayist's mind and character documenting his inmost convictions and sentiments on a wide range of topics. To read the text was to enter into an intimate and edifying communion with its author. The Essais were treated not as a complex or enigmatic text, but as a leisurely and companionable livre de chevet, a ‘livre de sagesse’ or moral handbook offering humane insight and aphoristic wisdom to an essentially docile reader. The text's form was taken to be incidental to the study of its contents, and subordinated to the important task of lending unity to Montaigne's disparate reflections by reconstructing his philosophical and moral views and situating his masterpiece within the wider context of his life, times, and personal development as a thinker.

This pedagogical and mimetic understanding of the text was subjected to devastating critique by Jean-Yves Pouilloux in his still influential polemic Lire les Essais de Montaigne, first published in 1969 and reissued in an expanded edition in 1995.Footnote ³ Pouilloux's thesis was that all attempts to establish Montaigne's philosophical ‘stance’ or to delineate his moral ‘teachings’ involve a radical censorship of the text, notably through the compilation of disparate quotations, lifted from their contexts and thus shorn both of their tentative and equivocal character and of their self-reflective and self-critical force. Anthologizing the Essais in this way, he argued, requires a staggering méconnaissance of some of Montaigne's most striking and distinctive statements about his work, in particular his disavowal of didactic authority, consistency, and certainty, and his self-referential elevation of ‘manière’ over ‘matière’, drawing the reader's attention away from the subject-matter in hand towards an inquiry into the manner or process of thinking and judging in itself.

Pouilloux's aim was not, however, to undermine the philosophical import of the text. On the contrary, he was committed to claiming that ‘Montaigne est penseur, les Essais sont un livre de “philosophie”’,Footnote ⁴ albeit in a highly unconventional sense. His point was rather that efforts to impose an overarching and coherent order on the text, against the grain of Montaigne's discontinuous and dissonant reflections, involve a refusal to ‘read’, a way of evading the intellectual challenge offered by the Essais and its irrepressible capacity for self-subversion. In this perspective, Montaigne does not have a philosophy, a politics, or an ethics to impart: the achievement of the Essais lies instead in its acknowledgement (and enactment) of the inconsistency and ambivalence of all philosophical discourse.

The advent of deconstruction from the late 1970s onwards drew further attention to the text's self-consuming and aporetic tendencies, undercutting traditional readings of the work as a transparent, ‘consubstantial’ witness to Montaigne's essential self and thought. In this post-structuralist light, the text performs its own inability to achieve either mimetic representation or wisdom, thwarting attempts to excavate its fundamental meanings and purposes by encompassing it within a coherent, unambiguous, ideological framework. This approach opened the way to a fertile rediscovery of the Essais as a text, as opposed to a disembodied system of ideas, highlighting Montaigne's writerly preoccupation with questions of imitation, representation, rhetoric, and textuality. The burden of interpretation was drawn away from the task of uncovering and understanding the broader philosophical, ethical, and political implications of the Essais, towards a more self-consciously literary exploration of the poetics of Montaigne's writing and its playful capacity for self-deconstruction, deferment, and duplicity.

The last fifteen years or so, by contrast, have witnessed an increasing appreciation of Montaigne's status as a philosopher.Footnote ⁵ Although perhaps most strikingly epitomized by The Cambridge companion to Montaigne (2005), edited by Ullrich Langer, which is part of a series on major philosophical thinkers, this development is exemplified by all the studies under review. Ann Hartle's Michel de Montaigne: accidental philosopher (2003) is ‘intended to show that Montaigne is a philosopher’, arguing that ‘although his Essays have always been acknowledged as the origin of a new literary genre, they have never been recognized as philosophical in the deepest sense’ (p. 1). In ‘Route par ailleurs’: le ‘nouveau langage’ des Essais (2006), André Tournon describes the Essais as ‘un ouvrage philosophique difficile, d'une complexité insolite’, to which a ‘lecture purement littéraire’, for all its merits, cannot always do justice on its own, appealing instead for a ‘nécessaire croisement des investigations littéraires et philosophiques’ (p. 386).Footnote ⁶ As Terence Cave puts it in his How to read Montaigne (2007), a short, introductory volume addressed to readers approaching the text for the first time, the Essais represent ‘probably the richest and most productive thought-experiment ever committed to paper’ (p. 3). This is the case despite the fact that philosophy features in the Essais as an ‘accidental, unpremeditated’ feature of a wider process of self-exploration and self-regulation: ‘Montaigne did not set out to be a philosopher, and if the Essais have a place in the history of philosophy, it is on the sidelines rather than in the mainstream’ (p. 45). But Montaigne's text is still ‘a book to think with, an intellectual resource still remarkably potent more than four hundred years after it was written’ (p. 5).

This emphasis on Montaigne's achievement as a thinker has gone hand in hand with a resurgence of interest in the more properly ethical resonance of the Essais.Footnote ⁷ Zahi Zalloua's Montaigne and the ethics of skepticism (2005) interprets Montaigne's dubitative and self-contesting practice of writing in terms of an ethical ‘concern for and openness towards the other’, leading to a ‘recognition of the other as an object of care or caritas’ (pp. 4, 3). Biancamaria Fontana's study of Montaigne's politics: authority and governance in the Essais (2008) sets out to redress the ‘lack of proportion between the elevated status of Montaigne the writer – established by a vast and ever growing stream of literary scholarship – and the uncertain reputation he enjoys as moralist, philosopher, and observer of the social and political reality of his time’ (p. 2). Finally, James J. Supple's Les Essais de Montaigne: méthode(s) et méthodologies (2000) aims to ‘répondre au défi de la critique moderne pour voir dans quelle mesure un ‘Montaigne moraliste’ pourrait résister aux assauts d'une approche qui privilégierait une rhétorique de la mobilité' (p. 423), concluding that ‘face aux attaques répétées de certains critiques modernes, la critique traditionaliste, et nommément celle qui est axée sur une lecture éthique des Essais, n'a nullement besoin de rester sur la défensive’ (p. 80).Footnote ⁸

Montaigne's decision to present his thoughts in the form of discontinuous and eclectic reflections marked by tension and doubt, rather than work them into a systematic and controlled argument, has become central to any credible account of his project. Despite this, the accounts described above embody contrasting and sometimes incommensurable approaches to reading and writing about the Essais. Hartle, Fontana, and Supple tend to focus on elements of substantive continuity and purposiveness in the text, using them to discern meaningful structures within Montaigne's unsystematic and undogmatic reflections. In this perspective, the Essais represent a highly idiosyncratic philosophical exercise, but one which can and should ultimately be accommodated within a wider framework of preoccupations, intentions, and convictions, as the expression of a powerful and distinctive body of thought. Cave, Tournon, and Zalloua, on the other hand, all remain committed to a much more radically self-contesting and performative conception of the Essais. These scholars take their cue from Pouilloux's conception of the text in terms of penser rather than pensée, as a book concerned with the activity of thinking itself rather than the construction of a system of thought. Montaigne's philosophical practice is read as a process of productive self-criticism taking place in and through writing. For these writers, the intellectual achievement of the Essais lies precisely in its willingness to subvert its own certainties, and in its efforts to realize and represent this new way of thinking in language.

A clearer sense of this distinction can be obtained by comparing Ann Hartle's Michel de Montaigne: accidental philosopher and André Tournon's ‘Route par ailleurs’: le ‘nouveau langage’ des Essais. In some respects, the two accounts echo each other closely. Both books make a strong case for taking Montaigne seriously as a thinker and for reading the Essais as a purposeful and fertile exercise in philosophy, by drawing attention to the logical trajectories of thought at work within the text. For Hartle and Tournon alike, the tensions and uncertainties exhibited by Montaigne's writing do not simply cancel each other out; on the contrary, the process of reflection draws fresh impetus and focus from the text's openness to digression and doubt. In sum, Montaigne's refusal to conform his thoughts to linear models of argumentation and persuasion represents a groundbreaking philosophical strategy, rather than a purely literary device, allowing constructive philosophical reflection to take place outside the disputed realm of dogmatic and didactic certainty.

These important parallels aside, a conceptual and discursive gulf opens up between the two studies, eloquently encapsulated by the quotations from the Essais evoked in their titles.Footnote ⁹ Hartle describes Montaigne's ‘accidental philosophy’ as a ‘circular dialectic’ based on the ‘reconciliation of opposites’ (p. 91) and the rediscovery of prereflective truths. This account of the philosophical dynamic of the text is fundamentally at odds with Tournon's characterization of the Essais as an unrelenting ‘hunt for knowledge’, propelled forward by unexpected, indirect, and seemingly anomalous itineraries of thought (‘routes par ailleurs’). Hartle sets out to capture the characteristic logic of Montaigne's thinking by explaining what he means by describing himself as an ‘accidental philosopher’. Her study offers a compelling portrait of Montaigne's ‘way of being’ (p. 13) and ‘way of proceeding’ (p. 18), drawing the Essais together as a unified and coherent conceptual object, indelibly stamped with the mark of Montaigne's distinctive identity as a thinker. Tournon's study, by contrast, operates through a series of extended close readings, guiding us step by step through individual chapters of the text in order to draw out their internal tensions and perturbations. His analysis of the ‘new language’ at work in the Essais fastens our attention on the detail of the writing itself, as the site of a ‘stratégie d’énonciation telle que la parole puisse se contester elle-même sans s'annuler'Footnote ¹⁰ (p. 27), highlighting the unstable and ambivalent qualities of the Essais as text.

For Hartle, ‘either Montaigne is a philosophically inconsistent and even incoherent thinker – that is, … he is not a philosopher at all – or a way must be found to go somehow beneath the philosophical chaos of the Essays and to locate Montaigne's distinct philosophical voice’ (p. 13). Her careful reading of the Essais confutes foundationalist assumptions about what counts as philosophy and what does not. But in her eagerness to prove that Montaigne ‘takes up the most fundamental philosophical questions in a profoundly original, comprehensive, and coherent way’ (p. 1), she is often too keen to resolve tensions and questions which Montaigne keeps in play, frictions which Tournon is extremely skilful at exposing.

‘Route par ailleurs’ dissects the essai as a thoroughly sceptical form of discourse, an ‘écriture pyrrhonienne’ (p. 33) that continually advertises its status as ‘zététique’, or unresolved, ever-renewed inquiry (p. 16): ‘les investigations conduisent à des acquis provisoires, à des étapes sur un trajet sans terme; et la synthèse est toujours différée, ce qui suffit à exclure les certitudes définitives’ (p. 27).Footnote ¹¹ This practice of relentless questioning relies on semantic ‘ruptures’ and ‘inflections’ which ‘dérangent le propos initial, bouleversent ses critères de validation, s'inscrivent en faux contre les conclusions auxquelles il paraissait conduire’ (p. 12).Footnote ¹²

Crucially, for Tournon, these pivotal shifts in the logical development of a chapter generate a dialogical rather than dialectical process of reflection. They reconfigure the chapter's line of sight, bringing its conditions of inquiry into focus and into question, disrupting its claim to know and understand in favour of ironic self-awareness. Tournon's account of this process in the case of ‘De la modération’ makes this feature of his argument particularly clear: ‘le texte porte en lui sa propre critique. Plus exactement, il se dédouble. Car il n'y a pas là, à proprement parler, une contradiction, et encore moins un progrès dialectique: les assertions antagonistes se disposent sur des plans différents … Le tout forme un jeu de métalangage’ (p. 38).Footnote ¹³ The resulting discourse cannot be described as incoherent or indecisive, but nor is it reducible to a single point of view. A ‘montage’ ‘polyphonique’ of ‘propos de degrés différents’ (p. 120), Montaigne's writing obeys the more supple and complex logic of the essai, offering itself not as the truth, but as a ‘parole véridique en ce qu'elle ne méconnaît pas ses risques d'erreur, ou d'enlisement dans les ornières qu'elle a tracées’ (p. 12).Footnote ¹⁴

Tournon makes clear that, if we want to understand the Essais, we need to pay attention to the detail of Montaigne's writing and its capacity to inhabit plural points of view. Montaigne's aim is not to clarify and consolidate his opinions, but to ‘essay’ problems from a variety of conflicting and overlapping perspectives, without confining himself to any particular standpoint, and without setting final limits to the process of inquiry. This emphasis on rupture, deviation, and the ways in which they are present even in the more ostensibly conventional and straightforward chapters of the text rightly situates Montaigne's reflections within their immediate argumentative contexts, shedding light both on the logical configuration of the chapter as a whole and on the way in which the meaning of a particular utterance is dependent on its place within it.

This almost exclusive focus, however, on the text's ‘dispersion’ (p. 387), on its ‘distortions régénératrices de concepts, configurations énigmatiques ou aporétiques, silences marqués, multiplication des perspectives, discordances’ (p. 310) leaves important questions unanswered.Footnote ¹⁵ What draws Montaigne to particular problems? What persistent habits of thought (and of language) inform his treatment of these questions? What are the arguments, concepts, and categories at his disposal? Tournon characterizes Montaigne's project in terms of a narrowly epistemological preoccupation with ‘la vérité, ou plus exactement le type de vérité qui peut subsister après les ravages opérés par le pyrrhonisme dans le champ des assertions objectives’ (p. 14),Footnote ¹⁶ to the relative exclusion of the substantive questions under discussion in the Essais. Yet as Tournon himself has put it elsewhere, Montaigne is ‘un écrivain qui ne se prétend pas dépositaire de vérités objectives, mais de convictions et de problèmes’.Footnote ¹⁷ Might it be possible to combine this exemplary attention to the Essais's ‘manière’ with a deeper consideration of its ‘matière’ – not in the justly discredited sense of definite answers or single points of view, but in the more dialogical and polyphonic sense of Montaigne's ‘convictions’ and ‘problèmes’?

Readers of the Essais tread a fine line between disentangling the manifold threads of meaning that make up Montaigne's text, and displacing the text in favour of a monolithic set of propositions and ideas. Montaigne's politics, by Biancamaria Fontana, is a case in point. On the one hand, the book sheds important light on a much neglected and misunderstood aspect of Montaigne's project, convincingly undermining Montaigne's conventional self-representation as a politically disengaged sage, withdrawing from public life and society in favour of the untarnished seclusion of the tower. Fontana's Montaigne is a visionary thinker outraged by the prevalence of cruelty, injustice, and violence at all levels of society, suspicious of the self-interested and ‘utilitarian’ logic of Machiavellian reason of state, and driven by an overriding concern to construct a more compassionate and peaceful form of human community, based on everyday, unheroic virtues and a sincere, spontaneous, and essentially Christian sociability. This picture is developed through a series of subtle and insightful chapters discussing Montaigne's approach to justice and the law, the renovation of virtue, religious toleration, the role of trust in political societies, and the nature of political experience.

This way of organizing Montaigne's dispersed ‘political’ reflections is deeply at odds, however, with the unsystematic and eclectic texture of his thinking. The book offers a ‘profile view’ of Montaigne as a political thinker, ‘focusing upon those aspects of his reflection that are relevant to the understanding of politics’ (p. 1). As Fontana acknowledges from the outset, this choice of perspective represents a conscious abstraction from the Essais themselves, which were clearly not designed as a contribution to political theory (pp. 1, 139–41). Her description of her project as an exercise in portraiture, albeit one executed from a particular angle, identifies Montaigne himself, his political career and thought, rather than the text per se, as the ultimate object of study. In this light, the Essais take their place alongside Montaigne's letters and other writings, as an archive offering access to the author's understanding and experience of politics. This opens the way towards an essentially biographical and conceptual rather than textual account of Montaigne's ‘contribution to politics’.

Montaigne's politics is at its strongest when it speaks of preoccupations and problems rather than of positions and answers. But it is still difficult to see how Fontana's approach can defend itself against the claim that the Essais do not allow us to reconstruct anything so determinate and clear-cut as ‘Montaigne's views on political issues’ (p. 141). As Zahi Zalloua puts it, invoking Pouilloux, ‘looking in the Essais for definitive answers – a single monological meaning – is to deny the skeptical or self-contesting nature of Montaigne's essayistic writing’ (p. 3). Zalloua accordingly analyses Montaigne's ‘ethics of care’ not by probing the substantive moral insights conveyed by the text, but by treating it as a feature of Montaigne's sceptical writing practice, through three chapters examining the representation of Socrates in ‘De la physiognomie’, of de la Boétie in ‘De l'amitié’, and of the Amerindians in ‘Des cannibales’ and ‘Des coches’. Zalloua is more particularly concerned with the way in which Montaigne's ‘art of prolonging’, through the introduction of digressions and allongeails, ‘alters’ his writing and thereby opens it to alterity, allowing him to ‘essay’ (rather than claim cognitive mastery over) the other. In Zalloua's reading, Montaigne's polyphonic and inconclusive prose knowingly highlights the other's resistance to apprehension, disrupting attempts to subordinate the other to a determinate, fixed interpretation, or to appropriate the other by reducing it to what is already known.

Zalloua's focus on the way in which Montaigne writes and rewrites the Essais is far removed from Fontana's interest in the overarching political vision that emerges from the work as a whole. In one sense, however, the two studies do have something in common: a tendency to present the text in terms that Montaigne himself would most likely not have recognized as an accurate description of his project. Zalloua sets out to bridge the divide between ‘poeticist’ approaches to Montaigne that ‘begin with the text’ and ‘contextualist’ approaches that ‘begin with history’ (p. 5), a rift diagnosed by Philippe Desan in an influential review article published in 1991.Footnote ¹⁸ In practice, however, his Levinasian reading of the Essais tends to sidestep the question of the text's historical identity, by bringing contemporary vocabularies, imperatives, and concerns to bear on Montaigne's project. The fusion of hermeneutics and ethics and the dialectic of self and other that lie at the very heart of Zalloua's study reflect a distinctively postmodern preoccupation with the unravelling of logocentric and autonomous models of subjectivity, placing it firmly outside the discursive worlds inhabited by the text.

It seems more productive to think of history not as something opposed to poetics or situated outside the text, but as an integral aspect of the text itself. As Terence Cave puts it in the introduction to his book, asking ‘whether we are to read the Essais primarily as a product of late Renaissance humanism, steeped in the cultural habits of that period, or as already a remarkably modern work’ is both unhelpful and unnecessary. In a striking echo of Montaigne's own account of the Essais as a series of ‘crotesques et corps monstrueux’,Footnote ¹⁹ Cave cites an example from Wittgenstein's Philosophical investigations (ii.xi), that of a trick picture which can be seen as the head of either a duck or a rabbit, but not both at the same time. ‘Any viable reading’ of the Essais, he urges, ‘will need to see both aspects, shuttling between them as between the duck and the rabbit’ (pp. 3–4).

This call for a more ecumenical and flexible critical practice, conferring complementary and countervailing approaches to the text, is echoed by James Supple – ‘à un livre pluriel convient, selon nous, une approche plurielle’ (p. 421)Footnote ²⁰ – and by the heterogeneous collection of essays commissioned for the Cambridge companion to Montaigne. Supple offers a remarkable, if at times austere, appraisal of recent critical approaches to Montaigne. His book begins not with a didactic statement of method, but with two antithetical readings of ‘De la gloire’, an ‘analyse traditionelle’ emphasizing its thematic and argumentative coherence, and an ‘analyse d'inspiration derridienne’ highlighting instead the chapter's self-consuming circularity (p. 15). As Supple explains, ‘il n'y a pas de recette magique qui puisse nous permettre de trouver sur les Essais ce qu'on a appelé “la perspective juste”’: to a large extent, ‘les résultats auxquels aboutissent les montaignistes sont le reflet direct de la méthodologie adoptée’ (p. 95).Footnote ²¹ The point is, however, that we have much to learn from these disagreements: carefully studied, debates over method allow us to ‘mieux définir la nature des problèmes que nous pose un genre qui met en cause nos stratégies de lecture’ (p. 96).Footnote ²²

The bulk of the book is accordingly taken up by studies of individual chapters of the Essais, as the starting-point for a rigorous and even-handed discussion of the advantages and disadvantages of different critical approaches, from psychoanalysis to riffaterrian semiotics. The outcome of this exercise is both to validate the contribution made to Montaigne studies by a revitalized ‘critique thématique’, sensitive to the context of Montaigne's utterances and in particular to the persuasive strategies that underpin them (and therefore far removed from the reductive practices castigated by Pouilloux), and to highlight more generally the need for productive dialogue (or at the very least peaceful coexistence) between divergent approaches to the text.

In a similar vein, the Cambridge companion offers less a comprehensive, unified survey of Montaigne's thought (if such a thing were possible) than a compelling juxtaposition of perspectives, essaying Montaigne – and the meanings of ‘philosophy’ – from a variety of angles. These include, for example, studies of Montaigne's attack on scholastic epistemology (Ian Maclean), and of his Lucretian ‘naturalism’ (George Hoffmann). But there is also a chapter on the presence of classical antiquity in Montaigne's work, not ‘as a set of abstract propositions or an inert corpus of knowledge, but as a body of writing within a body of writing, woven piecemeal into the texture and text of the Essays as part of the act of composition’ (John O'Brien, p. 54), and a chapter on the New World probing the ‘textual geography’ (p. 77) of the Essais, the text's own vanishing points and unchartered spaces, the tropic connections between oceanic exploration and introspective self-discovery (Tom Conley). Perhaps most interesting of all are those sections which explore Montaigne's rhetorical self-presentation in his text, whether as noble prudens versed in the techniques of statecraft (Francis Goyet) or as an aristocratic author that is his own patron, rather than a mere maker of books (Warren Boutcher).

But it is in Cave's How to read Montaigne that the plurality and historicity of the Essais are most lucidly expressed. Cave's presentation of the Essais as an essentially ‘cognitive’ rather than strictly philosophical exercise, aiming to provide ‘a documentary account of the mind's activities’ (p. 3), directs attention away from the vertiginous logic of sceptical self-contestation towards Montaigne's ‘fascination … with the epistemology of self-observation, and even more with the cognitive means by which the mind's workings may best be traced and delicately teased apart’ (pp. 94–5). Thinking of the Essais in these terms, as an index of ‘possible imaginings’ and ‘potential ideas’ (p. 64), allows Cave to emphasize Montaigne's familiarity and its strangeness in turn, to maintain what he calls the ‘tension of difference’, by reading the text as a liminal, early modern cultural artefact, suggestively close to our own conceptual and discursive habits, but still as foreign to us as the cannibals were to Montaigne (p. 4).

For Cave, as for Tournon, Montaigne's refusal to confer consistency, certainty and explicit order on his reflections represents an opportunity rather than an obstruction. In this perspective, it is just as unhelpful to assert that the Essais essentially defy interpretation, subverting all attempts to make sense of the work as a whole, as it is to claim that they offer unambiguous moral teachings or clear-cut lessons in philosophy. Montaigne's writing encourages us instead to be more searching and flexible as readers, alerting us at every turn to the importance of plural and open-ended interpretation.

II

If the challenge raised by Montaigne's writing is to read it in the undogmatic and elastic ways outlined above, to what extent might the new Pléiade edition be said to help us in this task?

The 2007 edition of Montaigne's text differs greatly in conception from the older Pléiade Essais edited by Albert Thibaudet and first published in 1934, which it will now replace.Footnote ²³ Thibaudet confined his notes to a minimum and presented the establishment of the text as an uncontroversial, essentially closed case. The apparatus of the 2007 edition, by contrast, takes up almost as many pages as the text and includes a full recension of Montaigne's reading notes, a new set of annotations for the Essais themselves, and up-to-date bibliographies for each chapter and for the work as a whole. This alone would suffice to qualify it as an important contribution to Montaigne studies.

The fundamental significance of this new edition, however, lies in its controversial presentation of a continuous, unlayered text, based on the first posthumous edition prepared by Marie de Gournay and published in 1595.Footnote ²⁴ This editorial strategy sets it in stark contrast both with the Villey/Saulnier edition of the Essais (VS),Footnote ²⁵ which has for many years been practically synonymous with Montaigne's text, and with André Tournon's Imprimerie Nationale edition (IN),Footnote ²⁶ VS's most convincing rival to date. Unlike Jean Céard's 2001 Livre de Poche edition, the first to reproduce the Gournay edition in its original form,Footnote ²⁷ the Pléiade project offers a genuinely critical edition of the 1595 text – a task that has not been undertaken since Courbet and Royer's now rare edition of 1872.Footnote ²⁸ This approach constitutes a welcome challenge to existing editorial practices, which have granted superior status to a version of the text unknown until the end of the eighteenth century and largely unread until the start of the twentieth, and which have consistently presented the text in a segmented and transitional form that Montaigne himself would not have recognized as his own.

In common with the overwhelming majority of twentieth-century editions, including the 1934 Pléiade text, VS and IN are based not on the 1595 edition, but on the so-called Bordeaux Copy (the Exemplaire de Bordeaux, EB), a copy of the 1588 edition containing extensive manuscript corrections and additions in Montaigne's hand.Footnote ²⁹ A small number of passages are present in EB but not in Gournay's edition, about 200 are to be found in the published text alone, and about 50 appear in different versions in each. The great merit of IN lay in its being the first edition of the Essais to offer a comprehensive and accurate transcription of EB, taking into account the thousands of adjustments made by Montaigne to the capitalization and punctuation of the 1588 text. These emendations had been completely ignored by VS and all other modern editions. But Tournon's dismissal of the posthumous edition as ‘un prestigieux apocryphe’Footnote ³⁰ aligned him firmly with twentieth-century editorial tradition, although in practice, like all of Montaigne's modern editors, Tournon was left with no choice but to appeal to Gournay's text to provide readings for those marginal additions to EB which were mutilated through careless cropping when it was rebound in the eighteenth century.Footnote ³¹

The Pléiade editors, by contrast, insist on identifying the full 1595 edition of the Essais with the final authentic state of Montaigne's text. This revisionist approach extends not only to the choice of copy-text, but also to the way in which that text is presented to the reader. Whereas both IN and VS break Montaigne's flowing prose into smaller, more manageable blocks of text,Footnote ³² the Pléiade edition follows Céard's lead by preserving the Essais's original layout.Footnote ³³ The result – a dense, continuous text interrupted only by chapter breaks and occasional, indented Latin poetic quotations – is less immediately accessible to a twenty-first-century audience than any of the other editions on the market, but is much more in keeping with the rambling and polyvalent character of Montaigne's writing. As Bernard Croquette and John O'Brien have recently emphasized, the original text's minimal paragraphing allows for multiple readings based on different ways of combining and dividing Montaigne's reflections. It is up to the reader to judge whether a given sentence offers a tacit extension or refutation of the sentences surrounding it, or to perceive the possible connections between two seemingly unrelated comments. By grouping sentences into paragraphs which, in the editor's view, deal with a common theme or correspond to a particular stage in an argument, VS and IN dramatically predetermine and curtail the signifying potential of the text.Footnote ³⁴

A more far-reaching but also more contentious feature of the Pléiade edition is its emphatic rejection of prevailing ‘genetic’ approaches to the text. Most twentieth-century editors, Tournon included, had tended to divide the Essais into distinct chronological strata, corresponding to passages present in the first edition of 1580, text added between 1580 and the 1588 edition, and manuscript additions to EB made between 1588 and Montaigne's death in 1592. In the case of VS, this took the form of capital letters (A, B, C) bracketed in the text; in IN, these indicators were relegated to the margins of the main text. For the Pléiade editors, by contrast, the only viable, consistent, and faithful approach to the Essais is to reproduce the final and most complete version of the text (in this case, the 1595 edition) without differentiating successive stages of composition, and to confine all variants to the end of the volume.

The case for the 1595 text, first of all, is strong. The argument is laid out in some detail both in the introductions to the Pléiade edition and in Philippe Desan's collection of essays, Montaigne dans tous ses états (2001), notably in the chapter entitled ‘L’exemplar et l'exemplaire de Bordeaux' (pp. 69–120). Drawing on a hypothesis first formulated in the 1970s by R. A. Sayce and David Maskell, these scholars argue that Gournay's text was based not on a corrupt transcription of EB, irrevocably contaminated by careless, misguided, or self-serving editorial interpolations, but on a second, more advanced set of authoritative revisions prepared by Montaigne before his death.Footnote ³⁵ According to this view, EB is not the final and finished draft of the Essais, ready to be sent to the printers or copied out by a scribe for this purpose, as Pierre Villey and others had assumed. Rather, it is an earlier, private working copy, superseded by one or more later copies faithfully transcribed in the 1595 edition. This narrative is supported both by the material appearance of EBFootnote ³⁶ and by what is known about the practical circumstances surrounding the preparation of the posthumous edition by Gournay and Montaigne's close friend Pierre de Brach.Footnote ³⁷ In this light, EB ought by rights to cede its place to the posthumous edition as a more mature and perfected version of the authorial text.

Two shifts in perspective have lent further credibility to this alternative account. First, the special status of EB has been considerably undermined by our increased awareness of the editorial history of the Essais from the sixteenth century to the present day. Desan's work in particular seeks to free readers from the mesmerizing materiality of EB as a rare autograph ‘manuscript’, by highlighting the contingent and anomalous character of its romantic and positivistic twentieth-century fetishization as the culmination of Montaigne's genius.Footnote ³⁸ EB's extraordinary prestige and hegemony since the early twentieth-century appears, in this light, not as the natural outcome of intrinsic philological qualities, but as the product of accidental historical circumstances and as an almost aberrant episode within the wider editorial history of the Essais.

Secondly, these efforts to normalize the 1595 text by appealing to its historical significance have coincided with a critical reappraisal of Marie de Gournay's role in preparing the posthumous edition. Traditional portrayals of Gournay as at best a naïve incompetent, and at worst a self-aggrandizing profaner of Montaigne's text, now appear to have been shaped by misogynistic prejudices rather than concrete textual evidence.Footnote ³⁹ Recent scholarship has tended to support a much more favourable characterization of Gournay as Montaigne's close associate during the final years of his life, as a literary figure in her own right, and as an editor of unquestionable competence and good faith.Footnote ⁴⁰ This picture is supported by Gournay's scrupulous efforts to correct errors in all extant copies of the 1595 edition by hand, and by her struggle to uphold the text's integrity and conformity to Montaigne's wishes, in preface after preface and in edition after edition, against the threats posed by pirated copies, commercial vulgarization, and attempted ideological appropriation, up until her death in 1645. Under such circumstances, it is arguably more reasonable to treat the 1595 variants as genuine authorial revisions, rather than to claim that Gournay wholly invented them when she had no obvious motivation for doing so.

There is something problematic, however, about the Pléiade edition's tendency to discount the difficulties attached to a text about which so little is known with certainty, except that it was prepared after the author's death and that it diverges considerably from a set of revisions, however preliminary, whose authenticity is by contrast beyond doubt. In the absence of convincing evidence to the contrary, it seems safe to assume that the Gournay edition was based on one or more alternative copies approved by Montaigne. The text published in 1595 cannot, however, be regarded as a simple, unmediated reflection of this lost source.

Gournay may, for instance, have felt compelled to perfect a process of comprehensive redrafting left unfinished at Montaigne's death by offering corrections and modifications based on her own judgement of what Montaigne had or might have intended, or indeed on verbal indications received from Montaigne himself. This hypothesis is arguably no more speculative and no less plausible than the one on which the Pléiade edition is based. Assuming, like the Pléiade editors, that we want an edition that reflects as closely as possible the agency of its named author, how are we to assess the autonomous decisions made by de Brach, Gournay, the publisher Abel L'Angelier, and his compositors in establishing and preparing the text for publication, decisions over which Montaigne had no final or direct control?

In the absence of the copy itself, as André Tournon continues to remind us, the 1595 edition will inevitably compare unfavourably with EB, the only fully accredited witness to Montaigne's post-1588 intentions. Tournon's reservations about the 1595 text focus on its failure to conform, in roughly 50 per cent of cases, to the revised capitalization and punctuation introduced by Montaigne in his manuscript corrections to EB. The significance of these apparently minor emendations is made clear by the autograph list of instructions placed at the start of EB, in which Montaigne specifically describes his writing as a staggered and sinewy ‘langage coupé’ characterized by frequent full-stops and upper-case letters. Tournon argues that these features of EB serve a philosophical as well as an aesthetic function, helping to emphasize moments of irony, tension, and paradox that disrupt the ordered flow of thought and frustrate the reader's desire for finality. As he puts it in ‘Route par ailleurs’, ‘l'achèvement compte moins que l'attaque qui lance le mouvement ou la reprise qui le fait rebondir et le prolonge; les clôtures sont provisoires, susceptibles d’être transgressées par une relance du propos … Ce n'est pas un style qui … est proposé, mais un mode de pensée autant que d'expression' (p. 378).Footnote ⁴¹ The 1595 edition's tendency to suppress these ‘arêtes vives du texte’, the living armature of the text (p. 379), has a profound impact on the way in which it is read, as Tournon shows in an appendix of about eighty exemplary passages.

The Pléiade editors state that ‘l’édition posthume retouche de façon plus fréquente et plus systématique le texte et en particulier la ponctuation de l'édition de 1588 que ne le fait l'Exemplaire de Bordeaux' (p. lii).Footnote ⁴² They claim that Gournay's text offers a more systematic realization of Montaigne's ‘langage coupé’ than EB, accentuating ‘l'effet d'un style coupé, nerveux et véhément, scandé avec solemnité par de nombreuses majuscules’ (p. xli).Footnote ⁴³ This assertion is difficult to verify, however, and it is hard to see how it can be reconciled with Tournon's findings. From the Pléiade editors' point of view, EB is little more than a rough draft, to which the 1595 text, based on a ‘mise au net’, a polished and perfected version, is manifestly superior. Tournon's case-by-case analysis turns this argument on its head. In all the instances he cites, it is the 1595 text that falls short of EB, not vice versa: shorn of EB's striking inflections in emphasis and tone, the posthumous text loses in sharpness, force, and sometimes even in logical clarity. It is hard to escape Tournon's conclusion that these infelicities represent editorial interpolations, caused by a failure to grasp the role of punctuation in structuring Montaigne's thought. But this claim has the effect of fundamentally undermining the Pléiade editors' appeal to the 1595 text as an essentially unmediated expression of Montaigne's final intentions.

The real problem with the Gournay edition, then, is that we know so little about the copy at its source and about the process leading from it to the 1595 published text. The case for its authenticity and the case against it both rely, to an unavoidable extent, on speculation: neither can be asserted with absolute confidence, and neither can preserve itself entirely from accusations of subjectivity and circularity. Scholars can no longer afford to ignore the 1595 text – this alone is ample justification for the Pléiade project. But the problems surrounding the posthumous edition show little sign of disappearing completely.

The choice of the 1595 edition as copy-text leaves the editors with no systematic means of identifying passages only to be found in that particular state of the text, even though these are the sections which are most subject to doubt. The 200 or so readings for which the 1595 text is the only source are fully endorsed by their positioning within the Pléiade edition's main text, whereas variants from EB are confined to the endnotes and thus excluded from the text proper. The problem is not that these decisions are necessarily ungrounded or mistaken, but that the uncompromising character of this arrangement leaves no space in which to acknowledge their relative uncertainty. There is no easy or immediate way for the more sceptical or cautious reader to know which passages are affected by this irreducible and unavoidable deficit of authority.

This ellipsis is of no consequence to readers who share the Pléiade editors' firm convictions about the authority of the 1595 text in all its parts, or indeed to readers whose primary interest is in the 1595 text as a historic edition, as the text which Pascal, Voltaire, and Rousseau read, regardless of its strict authorial status. But it does matter to those who are committed to the idea of a critical, authentic text, unwilling to dismiss the 1595 text out of hand, and yet reluctant to take its readings on trust. The new Pléiade edition will therefore remain problematic to the agnostic majority who may feel the need to make up their own minds about the passages involved, or who wish to read the 1595 text in full awareness of its limitations.

Ultimately, however, the Pléiade editors' choice of copy-text may prove to be less controversial than their radical rejection of the long-standing tradition of ‘genetic’ stratification. By dividing the Essais into three stages of development, Villey turned his edition into a circular exemplification of his own thesis about the ‘evolution’ of Montaigne's thought, from juvenile stoicism through sceptical crisis to epicurean philosophical maturity. ‘Courtes additions’ and ‘remaniements de texte, même prolongés, qui intéressent plus le travail du style que l'histoire de la pensée de l'auteur’ went unrecorded, simply because they did not contribute to this overarching narrative.Footnote ⁴⁴ Additions and modifications made between 1580 and the second edition of 1582 did not have any place within this tripartite schema either, and so were relegated to the footnotes. Moreover, because EB had been chosen as copy-text, text labelled ‘A’ corresponded not to the 1580 edition itself, but to the passage as it appeared in EB. There was thus no place in VS where a number of crucial variants could be recorded – for example, passages from earlier editions appearing in a revised form in EB, and corrections and amendments made within the manuscript sections of EB.

Some of these problems were resolved through the publication of IN, which distinguished text from the 1580 edition (A1) and text from the 1582 edition (A2), and which recorded in the endnotes those earlier rejected variants which did not find their way into EB's final state. However, diachronic approaches to the text still present a number of practical and conceptual difficulties. First, stratification offers at best an artificial approximation to the complex compositional history of the text. The way in which Montaigne revised the Essais – through both lengthy insertions and minute corrections – makes all attempts to represent its different available states extremely convoluted and baroque, forcing editors to choose between precision and clarity. Moreover, as Desan makes clear, even the most exhaustively layered text arbitrarily privileges and unifies a few stages of development over the countless corrections and additions for which no material evidence survives. Far from offering a complete picture of the transformation of the Essais over time, it abridges the process of continuous revision into a series of still frames, bringing together into one layer passages which may have been composed several years apart.Footnote ⁴⁵ By contrast, the Pléiade editors' decision to reproduce the 1595 text in its synchronic, published form, without any indication of chronological layers, allows them to offer a more comprehensive set of variants than any other existing edition of the Essais, through an apparatus incorporating readings from the 1580, 1582, and 1588 editions, passages only found in EB, and even postulating four different stages of composition and correction within the manuscript sections of EB.

A further objection to the genetic approach is that it is difficult to classify variants chronologically without engaging implicitly in a teleological ordering of the text. We may well reject Villey's evolutionary account of the Essais as a simplistic narrative of progress towards ever-increasing perfection. However, where certain passages or formulations are highlighted as later additions, we will probably feel inclined to privilege them automatically as more mature and more worthy of our interest than earlier segments. The reader's experience of the text cannot fail to be coloured by the knowledge that a particular sentence dates from the first edition of the Essais (even though it may in fact be the fruit of numerous, invisible, prior revisions), that a particular phrase was added to it ‘in 1588’ as an afterthought, and that the paragraph which follows it in EB constitutes Montaigne's last word on the subject (even though he may have wanted to rewrite it, had he had the opportunity to do so). The layered text is an editorial artifice, far removed from the unified work that Montaigne sought to present to his reader.

As far as the Pléiade editors are concerned, the compositional history of the Essais is simply not part of the text. With the exception of passages added in EB (which are identified in the notes as add. sur EB), the edition offers no way of determining the order in which the different passages making up the text were composed. The rejected variants are presented in a rather condensed typeface and layout, as endnotes called up by lower-case letters in the main text, alongside references to Montaigne's sources or page concordances to other editions of the Essais. If a reader wants to use these to reconstruct a past state of a given passage, this is of course possible but extremely laborious. These variants cannot be read as part of the text; nor are they intended to be.

The 2007 Essais are in many ways emblematic of the Bibliothèque de la Pléiade's shift away from the diffusion and consecration of ‘great works’, towards the production of scholarly editions that call the unity and transparency of such texts into question.Footnote ⁴⁶ Despite this, however, and in keeping with the Pléiade's continued self-presentation as a ‘bibliothèque de l'admiration’ (in the words of André Malraux), Montaigne's most recent editors remain powerfully aware of their obligations to the Essais as a cultural object and public literary monument. This leads them to privilege unity and finality over the more accidental, hesitant, and palimpsestic qualities of the text as it exists in its successive versions. In their eyes, the true text of the Essais just is its culminating state – a text which VS, IN, and other diachronic editions have scandalously suppressed. In this perspective, there are no other textual states, only rejected variants: to read the Essais is necessarily to read them in their most complete and mature form, uncluttered by all but the most indispensable editorial interventions.

These assumptions are deeply problematic. The Pléiade editors implicitly identify authorial achievement and ‘achèvement’ (completion), positing homogeneity and coherence as evidence of aesthetic and intellectual finality. In this light, the text appears to gravitate towards a natural, preordained point of equilibrium and perfection: ‘comme si à ses yeux la forme du livre avait atteint son point d’équilibre … comme si notre cavalier avait enfin et pour toujours trouvé son rythme et son assiette … Entre 1588 et 1592, Montaigne a senti que son oeuvre avait arrêté sa configuration définitive' (p. xxvii).Footnote ⁴⁷ Earlier, less perfect versions feature as rough drafts along this route towards the best text of the Essais.

This approach could conceivably be defended if we knew for certain what the Pléiade editors appear to assume: that the 1595 text offers the final – i.e. finished and complete – version of Montaigne's work. In reality, however, disputes over authenticity aside, the most that can be said of the posthumous edition is that it is the last known state of the text, based on the most chronologically advanced set of authorial revisions.Footnote ⁴⁸ Two basic observations are invoked by the Pléiade editors in support of their much more ambitious and contentious claims: the fact that no new chapters are added by Montaigne after 1588, and the addition of the definite article in the 1595 version of the book's title (Les Essais and not merely Essais). These two points certainly offer a valuable corrective to the text's seductive self-presentation as an infinitely open and unending ‘cornucopian’ discourse (in Terence Cave's influential formulation),Footnote ⁴⁹ poised ambiguously between boundless fertility and degenerate proliferation, its perpetual movement arrested only by death. However, neither the architectonic stability of the Essais after 1588 nor the 1595 title's implication of finality can suffice to demonstrate that it contains the complete and final draft of the text, capable of superseding all prior versions.

Moreover, even if we did know for certain that the 1595 text presents the complete and final state of the text as intended by Montaigne at the time of his death, this need not detract from the significance of earlier variants, including those rejected in this final version, since all of these reflect prior stages of textual development. These aspects of the text assume a fundamental importance for a reader who identifies the Essais precisely with the complete set of its versions and revisions, avoiding any prior attempt to privilege one state of the text over another. In this perspective, the text that is corrected or amended is no less important or interesting than that which comes to replace it. Although these need not be the only or even the first questions that readers ask of Montaigne's text, and although the full details of the process are forever lost to us, we may still want and need to explore the protracted compositional history of the Essais.

There is something paradoxical, then, about the Pléiade editors' criticism of genetic editing and its attempts to structure or contain the unstable and palimpsestic features of a text undergoing continual, largely invisible revision. The dangers of hierarchical reading and chronological foreshortening are present in a different but arguably more pronounced form in the new Pléiade text, which resolutely suppresses the text's temporality and earlier versions. To this extent, the decision to present the text in a synchronic and unstratified form is itself the result of a particular editorial rhetoric, positing texts as closed and finished artefacts, rather than as evolving and unfinished discursive performances.

III

To make sense of Montaigne's work is first and foremost to understand what it means to write digressively and dialogically. As we saw in part I, this means taking doubt and divergence seriously, instead of neutralizing them in favour of an underlying unity of purpose or of conviction. Furthermore, as our assessment of the new Pléiade edition has shown, it also means reading the text as a changing tapestry of reflections and revisions, rather than as a monumental, integral whole.

Crucially, however, the Essais represent more than a purely formal exercise in critical thinking: they address particular problems and questions in particular ways, acting within and upon prevailing systems of discourse. To explain the Essais, to explain Montaigne's manière, is thus not only to elucidate the sceptical and self-critical dynamic of his writing (and re-writing). It is also to understand the preoccupations, presuppositions, and interpretative categories that nourish and shape his fluid reflections – to reconstruct the complex horizons of understanding and expectation which it inhabits and brings into being. How are we to account for these important aspects of Montaigne's penser without reducing it to a fossilized system of pensée?

It is clear that we need to think of the Essais as a complex landscape of shifting intuitions, inclinations, and concerns, rather than as the expression of fixed assumptions and deeply held beliefs. This is not to deny that the Essais, for all their resistance to argumentative closure, exhibit significant ethical and philosophical continuities (Montaigne's condemnation of physical cruelty, for example, or his suspicion towards universal systems of explanation). It is ultimately more helpful, however, to think about interpretation as a way of explaining what makes the Essais into the particular text that it is, rather than as a way of reconstructing Montaigne's patterns of belief, however ‘accidental’ and unprogrammatic. The Essais are best thought of not as an archival imprint of Montaigne's thought, but as an exceptionally flexible exercise of judgement, allowing Montaigne to draw upon and confront contrasting argumentative and rhetorical strategies. To speak of dispositions, then, is to evoke a dense tissue of thinking habits inscribed in the text, made up of persistent preoccupations and anxieties, recurrent motifs and configurations, echoes between chapters, and traces left by other texts, rather than a fixed anchor of fundamental preferences or principles. These features of Montaigne's thinking are far removed from anything that could be extracted from the text, paraphrased, and elucidated in terms of its propositional meaning or conceptual coherence.

Ideas, arguments, and concepts are best understood not as entities existing independently of language, but as embedded within wider lexicons and rhetorics. To explain the text, in this sense, is to read Montaigne as a bricoleur, appropriating and refashioning pre-existing tropes, vocabularies, arguments, and other textual materials.Footnote ⁵⁰ This allows us to move beyond a stark contrast between form and contents, manière and matière, by reading Montaigne's ‘dispositions’ in discursive rather than purely cognitive terms, as habits of language as well as thought. Furthermore, by thinking of Montaigne as a bricoleur, reclaiming and reshaping the eclectic thinking materials at his disposal, we are able both to locate the Essais within these imaginative and discursive frameworks, and to reserve a primary role for his own negotiation, contestation, and transformation of these resources. This allows us to see the text in performative rather than referential terms, as part of a history of representations, a text among texts, rather than as a source providing access to (and determined by) a more material, extra-textual reality. Instead of treating the Essais as a synoptic witness to a broader culture, as a distillation of an overarching mental universe, we are led instead to emphasize the irreducible particularity of Montaigne's project, by explaining what makes his text into the unique cultural object that it is.

Above all, this approach allows us to dispense entirely with the assumption that the Essais can and should be explicated in terms of a single world-view, and that it is in moments of consonance that Montaigne's presence is most truthfully disclosed. Instead of scrutinizing the work for evidence of latent conceptual patterns, persisting in spite of its self-critical impulse, our task becomes one of analysing the Essais as a text in conversation both with other texts and with itself. In this perspective, questions of coherence and continuity lose much of their importance: the focus of attention shifts instead towards recomposing the supple and plural discourses deployed by the text.

This is not intended as a stipulation about the only, or even the most ‘suffisant’, way to read Montaigne, but rather as one possible way of conceiving our relationship to the Essais, one which equips us to think historically about the text without abridging or subduing its oblique and roaming character. The exercise of interpretation, like that of editing, is always left unfinished, but it is not for that matter arbitrary or futile. Our understanding of Montaigne is advanced not only by the ability to explain and analyse the text in new and persuasive ways, but also by the self-reflective insight that comes from questioning and accepting the limits of our interpretative assumptions – from doubt and uncertainty as well as from recognition. Like the rhetorical strategies deployed by the Essais themselves, the contrasting reading strategies deployed by Montaigne scholars compel us to acknowledge the text's opacity, its resistance to unravelling, as well as its generous openness to further elucidation.