Published in 1942 in occupied Paris, Recherches sur le symbolisme funéraire des Romains is a magisterial book of over 500 pages, 105 figures and 47 plates (with several images on each). It was completed late in life by the great Belgian classical archaeologist and historian of religions Franz Cumont (1868-1947). The book infuses the artworks it describes with a rich interpretative frame drawn on a lifetime’s command of ancient religion and its literary texts—stressing a complex and nuanced symbolism rooted in philosophical allegory (notably Stoicism and Neoplatonism), in religious beliefs and in paideia. Imagined as a crowning monument to his long researches into the symbolic meanings of ancient art, the book succumbed almost immediately (and still within its aged author's lifetime) to a punitive and comprehensive 30 page review that substantially damaged Cumont’s reputation, at any rate in the English-speaking world. This was written by the great Anglophone historian of religion, A.D Nock.1 Yet, as Nock admitted, Cumont’s ‘magnificent and monumental’ opus was ‘an event in our studies’; and in many ways it remains so. For this reason alone, it is welcome that Brepols have reissued the book; but all the more so because Janine and Jean-Charles Balty have provided a superb historiographical introduction—a brief monograph in its own right—of about 150 pages that discuss the author, the text and its reception. I have no intention here of re-reviewing Cumont’s book, but I will comment on the Baltys’ excellent work and on my own view of the place of Cumont’s scholarship.

The Baltys’ account of Cumont's book, buttressed by deep and careful research among his papers in the Belgian Academy at Rome—some of them, but by no means all, published by Corinne Bonnet2—is divided into five sections. These address the complex history of the book’s long gestation and final production in the difficult years of the War (pp. XV-XXXVII); the book's goals and limitations (XXXVII-XLII); its range of receptions—both contemporary to the time of publication and over the longue durée (an exceptionally complex historiographical project, extremely well carried out, pp. XLIII-CV); a long and thoughtful series of keys to the rehabilitation of Cumont's book (CV-CLVIII) and finally a brief conclusion. In the course of this the Baltys make many palpable hits against various fantasies promoted by Cumont's detractors—for instance, that he was a Catholic believer and apologist, which is simply untrue (p. XVI, n. 3 and XCIV-V).3

Their summary of the immediate reviews and responses is extremely interesting. As they emphasise, with the exception of Nock, the early reception of Cumont's long-awaited volume was strongly positive. They do not make this point explicitly, but one might add that the enthusiasm extended across the political spectrum from scholars close to the Nazi regime, like Gerhard Rodenwaldt, to those on the Christian left and in the Resistance, like Henri-Irénée Marrou. After its early approbation, some of the more discerning noticed the issue around which Nock had struck: Marrou commented that Cumont was ‘in effect excessive’, while Angelo Brelich noted the dependence on ‘erudite speculation’.4 Although the Baltys clearly do not agree with Nock, they give a good deal of space to a fair account of his review (L-LVI). The main driver of their discussion, however, is an attempt to rehabilitate some of Cumont's positions and the thrust of his work. I will return to this question at the end, but first let me do two things that in my view they do not sufficiently attempt: first, to situate an approach to the volume within the methodological investments of the author’s oeuvre, and second, in the intellectual context of the book' s genesis and publication.

Cumont’s methodology for the use of images in writing religious history is something deep in his personal career and goes back to his studies of Mithraism in the late 19th century, when he was still in his 20s. Second, the time during which Recherches was written and published is of compelling interest because the stakes of scholarship, scholarly allegiance and interpretation in relation to the ancestral past of Europe as a whole could not have been higher in the context of the worst war into which Europe had ever been plunged. I will comment on each of these questions in turn.

In 1975, in a fundamental paper that demolished the authority of Cumont in Mithraic studies—a field he had created and dominated since the 1890s—Richard Gordon demonstrated some fundamental flaws in Cumont's method.5 Working with archaeological material, but attempting to produce religious history in the context of the success of Christianity in the Roman Empire, Cumont imposed a ‘narrowly intellectual and theological conception of religion’, ‘an a priori assumption’, onto what was in effect ‘iconographic evidence’ (Gordon, 1975, 216). To enable his argument, Cumont upheld the strong theory that Roman Mithraism was a western form of Iranian Zoroastrianism (p. 216-7), for which his knowledge was dependent on a vast array of Magian and orthodox Zoroastrian texts (p.219).6 Gordon's concerns were to show the inadequacy of Cumont's assumptions to the task of understanding Roman Mithraism, rather than to reveal Cumont's archaeological method per se. But several issues may be noted. First Cumont's commitment to matters of religion meant that he trod a rare route within the archaeology of his time—moving away from the stylistic formalism that dominated the German-speaking establishment, and especially the Viennese from the work of Alois Riegl in the 1890s to the second Vienna school, best represented in Classical archaeology by the deep formalist structuralism of Guido Kaschnitz-Weinberg in the 1920s and 30s.7 Rather, Cumont significantly anticipated the kind of art historical placement of forms, types and iconographies within a literary contextual frame designed to reveal underlying Weltanschauung that characterised the great Hamburg project of the scholars gathered at the Bibliothek Warburg, notably Erwin Panofsky, who specifically emphasised meaning over style, and Fritz Saxl, the Warburg Institute's director.8 Nothing in fact demonstrates the attraction of Cumont's model of working to the Warburgians so much as the most important forgotten book on Mithras of the last century—Saxl’s 1931 monograph, published at the apogee of the Warburg’s Weimar period, with its focus on the iconography of the bull-slaying scene in its form, symbolism, religious meanings and ultimately influence on Christianity.9

The method of Cumont's work on religious imagery was thus not the normative formalism of classical archaeology but a close conflation of image and text on models that were both greatly influential on the Warburgians and rather similar to them.10 However, where the kind of iconology practised by Panofsky, especially in the 1930s and later, could draw on a rich range of texts, close in date of production or reception as well as in place of production, to the images he chose to work on in the Italian or Netherlandish Renaissance or the Middle Ages, Cumont was forced to find ‘relevant’ texts wherever he could. In the case of Mithraism, this not only meant going outside the Roman world into Persia, but it also involved a ‘reconstruction of Mithraism culled from an enormous variety of sources of Iranian origin’ (Gordon, 1975, 243). It is the turn to a similarly enormous variety of texts, borrowed from anywhere and everywhere, to provide an apparently coherent explanatory frame for funerary symbolism that precisely defines the nature of Recherches, although it was published more than 40 years after the work on Mithras. It represents the same ‘Warburgian’ method applied to a different and considerably less coherent body of material. It is the consistent and unjustified a priori assumption of deeper meanings, allegory and symbolism underlying all funerary art combined with a hopeless lack of rigour or methodological clarity in offering any form of objective criterion for what should or should not count as an acceptable text to adduce in relation to a given monument, that is the real—and entirely justified—target of Nock’s demolition.11

But an analytic and brutal deconstruction of Cumont's method (whose weaknesses were not obvious to most readers at the time of publication12) needs tempering by historical contextualization. We must ask what was Cumont trying to do in Recherches and how was his method serving that purpose. Here, the issue is complex, fraught, but extremely interesting because of the period when the book was written. Let me approach it obliquely, by means of the Warburgian comparison. In 1933, with the advent of Hitler's regime in Germany, the Warburg Institute moved from Hamburg to London. It is a striking mark of the political frame governing the scholarly output of many who were associated with Warburg (most of them Jewish) that their art-historical scholarship during the Nazi period turned towards issues of Neoplatonism, symbolism and polysemy. There may be many explanations for this, but I would hazard that the positive politics underlying this line was a direct if subtle riposte to the essentialist and racial model of single meanings that underlay forms in the new, simplified and racist art history that took over the German universities in the Third Reich. To my knowledge, no one has explored a politics to Cumont's art-historical writings, but it seems to me that the impulse beneath the flurry of Neoplatonic polysemy that dominates Recherches, although certainly anticipated by Cumont's earlier work, belongs in its 1940s context, to the same instinct as the symbolic forms borrowed by Panofsky from Ernst Cassirer to be the underlying stratum of meaning in his Iconology, and the play of symbols in the works of Ernst Gombrich and Edgar Wind.13

In later years, notably the new art history of the 1980s, what had become a dogmatic application of textual sources to elucidate underlying meanings became viciously attacked as the discipline’s (then) youngsters turned on the dominant (Warburgian) voices of the post-war era. Something similar was at play in Gordon's attack on C, governed by the Oedipal necessity of killing the father. But we should not be distracted by the positivist call for getting closer to the real nature of Mithraism, funerary symbolism or whatever, in terms that are ultimately more meaningful or relevant to the discursive preferences of a later generation. In the era between 1939 and 1945, in the depths of a Europe condemned to a perpetual mortuary lament over the demolition of what was good about its own past and with little hope for any kind of future, Cumont's symbolic model of the imagery of the Roman tomb, as looking to a better place and one not reductive or brutal in its meanings, was a powerful personal statement about the present. It is this force, this politics, that Nock—writing from America and writing after the war—did not grasp, although his review came so swiftly after the initial publication.

Within the darkness of death, Cumont saw evocations and allegories of a better world and of salvation. In search of ‘revelation beneath the forms of obscurity’ and ‘to divine their secret’ (p. 13), he sought the aid of what in 1942 were forbidden masters—‘Philon le Juif’ (e.g. 13 and 16) and the intimations of what were considered lower races (‘des Syriens, des Indiens’, p.17). Is it pure chance that he chose to end the entire project with an appendix (broadly republishing an article 1916) on ‘un fragment de sarcophage judéo-païen’ (pp. 462-76, my italics on the key word)? Is it with pietas towards the Minister of Education in Pétain’s Vichy, or with a certain edge, that he repeatedly mentions the ‘Pythagorean’ basilica beneath the Porta Maggiore in Rome, whose major publication was the ‘brilliant volume’ by Jérôme Carcopino (p. 20, n. 59)?14 It is impossible for us now to grasp complex ambivalences and gestures in the politics of this writing—but my hunch is that a liberal resistance to totalitarianism and its collaborators was not entirely absent, in the call for a mystical interiority that could deny the intimations of the outer world.

It should be added—and it is perhaps unfortunate that this is not raised in the Baltys’ discussion—that Recherches was not Cumont's final statement on funerary symbolism, but that he continued to work on the subject through the later years of the War to produce the text of his last book, Lux Perpetua, which was published in 1949, two years after his death. Something of the politics of Cumont's late project here may be clearer from his own address at the sixth International Congress of the history of religions in Brussels in 1935, when the darkness was already closing on Europe, although it had not yet given way to war: ‘There are those who pose in panegyrics of violence and who vaunt the benefits of brutality, as the sole creative force of new societies; they seek an immediate remedy through physical coercion for the disorders we suffer. But what creations are more powerful and more durable than those of the spiritual forces, which have transformed peoples and overturned empires, like the invisible force of the wind that bends and uproots the forests? To totalitarian states that pretend to subject to their domination not only the actions but even the inner feelings of individuals, religious experience teaches how intimate convictions, pursued within external forms, offer an inviolable space of asylum in our heart of hearts. In this time, when all these nationalisms are so exacerbated, the development of religion shows us how a community of beliefs, once those of tribes or clans, then those of cities and nations, may aspire ultimately to universal significance and to create from heterogeneous and remote populations ties of allegiance more powerful than those of neighbouring kinship or blood…’15 It may be that today's ear is deaf to some of the notes of resistance to Fascist totalitarianism implied in Cumont's turn to personal faith and allegory. But his Recherches belongs to a moment when ‘external forms’ within the context of a ‘totalitarian state’ certainly ‘pretended to subject both actions and inner feelings to its dominion’ and when only an inner integrity, ‘the inviolable space of asylum in our heart of hearts’ could stand against the ‘violence’, ‘brutality’ and ‘physical coercion’. Arguably, Nock’s ear was already deaf to these issues in 1946. But for any fair assessment of Cumont's larger enterprise in Recherches, these themes cannot be ignored.

The impasse about how to make objects mean was not a new one in the argument between Cumont and Nock, nor has it been resolved since the 1940s. My aim here has been to show that the issues—in the funereal darkness of 1942— were very different from how they were seen in 1946 and from where scholars might go today; but they were not less profound. If we turn from historicizing our authorities to the philosophical and substantial point of how to write the history of religious art, the problem remains potent. If we are to go the Warburgian route of elucidating visual materials with textual supplements, we still need an objective criterion for knowing what texts we can reasonably apply and how to do this; for all their brilliant art historical verve, not one of the great scholars associated with the Warburg Institute has ever provided such a thing and nor did C. If we dismiss the range of potential textual material, and turn to the objects themselves to supply meanings, we may land up by simply forcing what passes for common sense in any particular moment (our own contemporary set of prejudices, usually) onto the images of the past in order to explain them. That is the risk of Nock’s proposal of ‘classicism and culture’ (Nock, 1946, 166) as his solution to resist the ‘maximizing’ impulses (p. 170) of symbolic and religious resonance that governed Cumont's interpretive vision at the darkest time of European catastrophe. I suspect the Baltys cannot be right in any attempt to reinvigorate Cumont's own interpretations, but there is no doubt—in the particular arena of the study of Roman sarcophagi, for example—that too much of the baby that Cumont so deeply cherished has been thrown out with the bathwater of his (in my view understandable and politically laudable) methodological excess.