After a long (writing-deadline induced) delay, I return to the blog with a book review. Though I have been busy writing this last month, I have also been reading the latest Gance-related publication. This handsome 300-page paperback is designed to accompany the forthcoming Cinémathèque française restoration of Napoléon. Having spent much of my adult life researching and writing about Gance (and Napoléon in particular), I am of course immeasurably excited about this new edition of the film. I will be attending the premiere of its presentation in Paris in July, so this book is a tremendously tasty preview of what to expect.
Firstly, the book is a lovely thing to hold and flick through. Though it is a paperback, it also comes with a dustjacket – a slightly odd combination that tends to be a little slippery to hold. That said, it’s filled with full colour reproductions of stills, portraits, posters, and – most of all – images from the film itself. The text occupies the bottom part of each page, while the top boasts a frame from the new restoration. Page by page, these frames cover the entire chronology of the film – including several fold-out spreads for the final triptych scenes. The text of the books contains nine essays that cover the film’s restoration, history, and cultural importance. Rather than go through them all in detail, I will group them into strands that discuss the film’s restoration, the new musical score, and the film’s genesis and ideology.
The restoration is the focus of pieces by Costa-Gavras (“La Cinémathèque française: une longue fidélité à Abel Gance et à son Napoléon”) and Georges Mourier (“L’éternel retour d’une restauration”). The former is the president of the Cinémathèque française and, as his title suggests, is both a history – and a kind of defence – of the institution’s relationship with Gance. Costa-Garvas traces the awkward history of the film’s restorations and the need for a more comprehensive attempt to reproduce the “Grande Version” envisaged by Gance at the end of 1927 (more on this later). As well as paying tribute to the previous versions assembled by Marie Epstein and Kevin Brownlow, Costa-Gavras also acknowledges the huge number of archives, funders, and cultural institutions that have collaborated for the new restoration. Of particular significance is his credit to the SACD (Société des Auteurs et Compositeurs Dramatiques) for confirming the world rights of the Cinémathèque française in relation to Napoléon (38). Anyone familiar with the film’s complex legal history will know that the rights to Napoléon outside France and the UK have always been claimed by another party. (A fact that is never mentioned in this book.)
As the head of the restoration team, Mourier has been working on Napoléon for nearly twenty years – and his passion for Gance long predates this project. His piece goes into more detail about the restoration process, though it cites (and reuses much information from) a more in-depth piece Mourier wrote some twelve years ago (“La Comète Napoléon”, Journal of Film Preservation, no. 86 (2012): 35-52). I mention this because both Costa-Gavras and Mourier summarize the principal versions of the film in a way that is not always the clearest exposition of the numerous different editions (and restorations). This relates to the way the new restoration has been advertised, both throughout this book and more generally in the press, which also merits discussion.
To do so, it’s necessary to recap the most important versions of Napoléon successively prepared by Gance in 1927. (The following figures are from Mourier’s 2012 article.) First, the “Opéra version” was shown in a single screening at the Paris Opéra in April 1927. This included two triptych sequences (the “Double Tempest” and the “Entry into Italy”) and measured 5200m. At 20fps, this would be approximately 225 minutes. Second, the “Apollo version”, which was shown over two days at the Apollo cinema in Paris in May 1927. This version did not include any triptychs and measured 13,261m (c.575 minutes), reduced to 12,961m (c.562 minutes) for release. Third, a reduced version of the Apollo edition that was prepared for release in America in November 1927 but never screened (it was subsequently butchered by MGM). It is this version that Mourier – and the entire 2024 book of essays – refers to as “la Grande Version”. It was not called “la Grande Version” in Mourier’s 2012 article, so this seems to have been rebaptized in the intervening years. Mourier has recently cited Gance himself as a source for this epithet – but provides no source as to where or when it was originally used. Furthermore, as noted by Kevin Brownlow (“Napoleon”, Abel Gance’s Classic Film (Photoplay, 2004), 146n) and Norman King (Abel Gance: A Politics of Spectacle (BFI, 1984), 148-9n), the version prepared for MGM in November 1927 included both the triptych and the single-screen versions of the Double Tempest and Entry into Italy (i.e. it could be shown with one, both, or neither of the triptychs, according to the requirements of exhibitors). Taking this additional/alternative material into account, Mourier (in 2012) gives the length of this version as 9600m (c.416 minutes, i.e. just less than seven hours). This figure is oddly absent from the 2024 book, as is the issue of how closely the new restoration relates to it.
Mourier’s contribution to the new book states that the Apollo version was 12,800m, “9 hours”, but states only that the “Grande Version” was “about 7 hours with triptychs” (225). Both these figures, and those used in Mourier’s 2012 article, assume a projection speed of 20fps. However, as detailed on the Cinémathèque française website, the new restoration runs at 18fps. (Brownlow’s restoration uses 18fps for the Brienne prologue, but 20fps for the remainder of the film.) Since the 2024 book and all the press reports use only runtimes (not length in metres), there is a pervasive confusion between the different versions of the film. The new restoration has a runtime of 425 minutes, which equates to approximately 8830m at 18fps. It is therefore somewhat shorter than the 9600m of the “Grande Version” (if we are to keep calling it that) as assembled by Gance in November 1927. (About 37 minutes shorter, at 18fps.)
The book also leaves unclear the precise method by which the contents of the Grande Version have been distinguished from the contents of the Apollo version. (Or even if this distinction was the goal of the restoration.) This is important, since the Grande Version was a reduced version of the Apollo version – and derived from the same negative. Mourier refers to a document he nicknames his “Rosetta Stone” in restoring Napoléon (236-7). It consists of a scene-by-scene breakdown of the Apollo version (divided into 36 reels), with length in metres for each sequence. Costa-Gavras writes that this document was “rediscovered in 2012” (35). But I presume it is the same document seen by Brownlow during his restoration, also discussed by Norman King (Abel Gance, 148-9). (I too went through it for my research in 2009.) Using this document to reconstruct the Apollo version is an obvious step, but was there a way of distinguishing footage that was used in the Apollo version but subsequently excised for the Grande Version?
This question is not addressed in any of the 2024 essays, nor in Mourier’s 2012 article. It is the same issue that arose with François Ede’s restoration of Gance’s La Roue. In that case, their blueprint for the restoration was the version released in February 1923. This was a shortened version of the premiere version seen in December 1922. Though Ede was unable to find all the footage from the 1923 version, he did find material from the 1922 version that he knew was subsequently excised. He therefore did not incorporate it into his restoration. (These few scenes are included in the extras on the DVD/Blu-ray release of the 2019 restoration.) Is there anything that Mourier has excluded from his reconstruction of the Grande Version, knowing that it was only used in the longer Apollo version? Or was all surviving material from the Apollo version used, regardless of whether it could be established to have been included in the Grande Version?
I also find it surprising that the 2024 book gives no runtime or physical length for the Cinémathèque française restoration, nor is the projection speed of any version given. This creates a false equivalence among previous restorations. Costa-Gavras, for example, records the temporal length of Brownlow’s restorations to compare them with that of the Cinémathèque française – but crucially does not mention the different projection speeds (32-4). Brownlow’s most recent restoration runs to 332 minutes, equating to 7542m. But while Costa-Gavras gives the impression this is 90 minutes shorter than the Cinémathèque française restoration, the divergent speeds means that the actual difference is only an hour.
Similarly, it is unclear to me why the 2024 book makes no reference to the given length (9600m) of the Grande Version. Only Dimitri Vezyroglou’s piece cites this figure, but he does not refer to it as the “Grande Version”. He states that this 9600m version was prepared in November 1927 for release in France, but for various reasons was not ultimately distributed in the form that Gance envisaged (115). Per all the other essays, Vezyroglou describes this version as “7 hours” – which (again) is only true with a projection speed of 20fps. Is there some doubt about the exact length or contents of the “Grande Version”? In which case, why insistently use this label to describe the new restoration?
I am also curious about the fate of the Double Tempest sequence. In his 2012 article, Mourier discussed elements that are known to survive from this triptych – and even provides a reconstructed triptych panel for one section (see below). However, the 2024 book makes almost no mention of it in relation to the new restoration or the decisions that led to it taking the form that it has. Joël Daire comments only that it “remains lost” (77), but Mourier never explains why or how it has been impossible to reconstruct – or the reasons why he chose not to attempt to do so. Given that it was an (optional) part of the version Gance prepared in late 1927, any decision to exclude it is also (necessarily) a creative one.
All the above relates to the main absence from the 2024 book (and, more generally, the information released by the Cinémathèque française): a discussion of the creative decision making involved in this restoration. The contributors acknowledge the sheer variety of (historical) versions and (modern) restorations of Napoléon, but the purpose of the book is ultimately to promote the singular (and presumably “definitive”) version of the film that the Cinémathèque française has prepared for worldwide release. While always paying tribute to earlier restorers (especially Epstein and Brownlow), the aura of definitiveness about the Cinémathèque française project carries a certain (unspoken) sense that the work of amateurs has now made way for the work of professionals. Brownlow’s history of his restoration of Napoléon is filled with personal anecdotes – his meetings with Gance, his obsessive hunt for material from the film, his taping together pieces of filmstock or sneaking behind Marie Epstein’s back to examine rusty tins of celluloid. In 2024, Frédéric Bonnaud writes that Brownlow’s account now “reads like a suspense novel” (58). I’m not sure if this is intended as a compliment or a criticism. It certainly contrasts with the way Mourier talks about the restoration process. In 2012, he described his work not as detection and intuition but as a scientific process of “geological drilling”: a combination of “vertical” and “horizontal” investigations to trace both the history of the film’s negatives and the multiplication of positive copies. The 2024 book expands this into a much wider discussion of the film’s history, but there is also an odd sense that the history of multiple versions has now come to an end: numerous paths have led to this single destination. But the staggering thoroughness of the Cinémathèque française project, and the wealth of primary documentation consulted, does not mean that there have been no creative choices involved – alternative paths not taken. Would (or should) a reconstruction of the “Grande Version” preclude the incorporation of any additional material from the original, longer Apollo version? Why choose 18fps rather than 20fps as the projection speed? Why not attempt to reconstruct the Double Tempest triptych?
Though these questions are specific to Napoléon, the archival and textual issues they raise are inevitable in any silent film restoration. Whatever the answers, it should be remembered that the ultimate goal of restoration is, after all, for the film to be shown to new audiences. Regardless of how the 2024 version relates to those of 1927, the Cinémathèque française can justifiably regard their restoration as the most satisfying presentation of Napoléon that can be achieved with the material they possess. Even if I remain unsure how the new restoration can claim to be “la Grande Version”, it is undoubtedly “une grande version” of Napoléon.
The music is perhaps the most significant aspect of creative choice involved for the presentation of the new restoration. In the 2024 book, the score is mentioned by several authors, but is the special subject of a piece by the composer Simon Cloquet-Lafollye, who compiled the new score to accompany Napoléon. While Cloquet-Lafollye never discusses previous scores (though the anonymous preface to his essay does (249)), other contributors cover the history of musical presentation – if not in much detail.
In his piece “Un film plutôt que sa légende”, Frédéric Bonnaud raises the fact that Napoléon was first seen in April 1927 with a musical score compiled by Arthur Honegger. For this, Honegger wrote a small amount of original music and otherwise relied on music from the existing repertory (including, in all likelihood, his own other compositions). But the difficulties of preparing both the film and the score for the premiere meant that the music was inadequate for exhibition, satisfying neither Gance nor Honegger. The performance in April 1927 (and, as I wrote elsewhere, in the Netherlands in August 1927) was something of a shambles. Thus, Bonnard rather breezily dismisses the composer’s involvement in the film: “So, no, dear friend, Arthur Honegger did not write the music for Napoléon” (44).
Since an earlier restoration of Napoléon presented by the Cinémathèque française in 1992 included a score based on the work of Honegger, compiled and expanded by Marius Constant, this attitude marks something of a shift. To highlight the inadequacies of Honegger’s music for the Opéra version is understandable, but to exclude his music entirely from the new score is a bold decision – especially considering the 2019 restoration of La Roue, where the Paul Fosse/Honegger score of 1923 plays such a pivotal role. In that restoration (which I discussed here), Honegger likewise wrote only a small percentage of the overall score – and much of this original material remains lost. Yet the musical reconstruction took Honegger’s involvement seriously enough to create new sections of music based on the material that does survive. Back in 2019, I was also told by the German music team responsible for the La Roue score that they had made some interesting archival discoveries relating to Honegger’s work for Napoléon. This was an avenue not pursued for the new Cinémathèque française restoration.
From his comments, Cloquet-Lafollye’s contract seems to have precluded any attempt to amend/expand Honegger’s surviving music for Napoléon to match the new restoration. (Until the BFI’s restoration of Napoléon in 2016, Carl Davis’s score included one of Honegger’s cues. This was Honegger’s counterpoint arrangement of “Le Chant du départ” and “La Marseillaise” in the final triptych. It worked well, and I do regret that it was replaced in 2016 with new music by Davis. Not that I don’t like Davis’s cue for this sequence, but it was a nice tribute to Honegger to at least preserve something of his music for modern presentations of the film.) Was it also impossible for Cloquet-Lafollye to include any of Honegger’s music from this period in the score? Why make room for Penderecki but not for Honegger? (Penderecki’s Third Symphony (1988-95), used by Cloquet-Lafollye, seems to me a rather undistinguished piece to choose in relation to almost anything else he could have picked from Honegger’s oeuvre, or from any other early twentieth-century modernist.)
Though such questions – no doubt involving copyright issues – go unanswered in the 2024 book, Cloquet-Lafollye at least discusses something of his methodology. He writes that there was “no question of creating a musical pastiche of the eighteenth century” (253), though he does cite work by Haydn and Mozart. He also wanted to avoid creating an unnaturally precise evocation of sounds on screen, for example gunfire: “Gance didn’t have the possibility of employing [such synchronization], so there was no question of my doing so” (252). His goal was to produce “a homogenous, coherent piece, in perfect harmonic synchronization with the rhythm imposed by the images” (256). Yet the very idea of “a homogenous, coherent piece, in perfect harmonic synchronization with the rhythm imposed by the images” would have been just as impossible for Gance to achieve in 1927 as the kind of synchronized sound effects that Cloquet-Lafollye shuns.
The 2024 book usefully lists all the pieces used by Cloquet-Lafollye in his score (303). Given comments I had read earlier by the composer, I was (happily) surprised to see so much music from the mid or early nineteenth century (some Beethoven, Schubert, Mendelssohn, and plenty of Liszt). I was also curious to see similarities between the music chosen by Cloquet-Lafollye for Napoléon and the music chosen by Fosse/Honegger for La Roue (and used in the restoration of that film in 2019). Both scores feature work by Dupont, D’Indy, Gaubert, Godard, Magnard, Massenet, Ropartz, P. Scharwenka, Schmitt, Sibelius, and de la Tombelle. Indeed, some of the same works used in La Roue – Gabriel Dupont’s Les Heures dolentes, Philipp Scharwenka’s Fantaisie dramatique, Fernand de la Tombelle’s Impressions matinales – are used by Cloquet-Lafollye for Napoléon. Though Cloquet-Lafollye’s essay makes no reference to this connection, these choices can hardly be a coincidence. Given that Honegger was involved in selecting the music for La Roue, it’s quite a backhanded compliment for Cloquet-Lafollye to use music that Honegger knew and used but not the music of Honegger himself.
I am left wondering what was the exact remit for Cloquet-Lafollye’s choices? The score is not an attempt to recreate the soundworld of Napoleon, nor the soundworld of Gance or Honegger. He has chosen to avoid too much synchronization, but conversely choses to create a “perfect” match for the rhythm of the images. Some of his other musical choices strike me as asserting a kind of retrospective cultural kinship between the film and like-minded (or like-spirited) music. Hence the inclusion of works with impeccable modernist pedigree – Mahler, Shostakovich, Webern – but which are also some of the music least likely to have been used for any screening in 1927. (Shostakovich had yet to write either his ninth or thirteenth symphony (from 1945 and 1962, respectively); the pieces by Webern and Mahler were not widely known outside Vienna.) But the like-mindedness (or like-spiritedness) is also a matter of creative interpretation. Whether Gance’s film is constitutionally “romantic” or “modernist” is a topic I have written about many times elsewhere, and it’s an issue that tends to come to the fore whenever music is discussed. Personally, I consider Napoléon a work of romantic imagination – and that this is the very source of the film’s modernity, fuelling its rich, strange, and profound inventiveness. Overlaying Gance’s astonishingly beautiful, often highly romantic imagery with layers of angst-ridden musical modernism does not always produce the best results.
But at this point, I am overstepping the bounds of this piece, which is (I remind myself) supposed to be a book review. I must see Napoléon with the new score before I judge how it works. All the extracts I have seen work very well, so I am not complaining about the use of the music – just querying the stated rationale of its compilation. I am very curious to see how Cloquet-Lafollye employs his wide-ranging musical choices.
The film’s genesis and ideology are discussed in pieces by Joël Daire (“Histoire d’une réalisation hors norme”), Dimitri Vezyoglou (“La circulation de Napoléon juqu’à la fin des années 1920”), and Elodie Tamayo (“Un cinema d’Apocalypse”). I fear I do not have the space to adequately explore these fascinating essays. What I would observe is that these are by far the most rigorous (and well-footnoted) sections of the book. Daire’s piece traces the pre-history of the film’s conception, especially the cinematic (and cultural) influences that shaped Gance’s imagination in the 1910-20s. It’s great to see the influence of American cinema – not just Griffith and DeMille but Fairbanks – being acknowledged (64-6), as well as Gance’s ambition to create a mode of world cinema (not simply a national one). In his contribution, Vezyroglou details the “tragedy” of Napoléon’s botched distribution within and beyond France. Much of this has been covered in Brownlow’s book, but Vezyroglou brings more archival sources to the story and enriches his account with more detail than many previous accounts.
Tamayo’s piece is the most interpretive (and imaginative) of the three, creating a marvellous picture of the film as “an apocalyptic poem, a work that demands cinema destroy the world in order to create anew” (135). Through its mission “to reveal an art of the future” (137), Napoléon sought to explode the spectator’s conception of spatial and temporal reality – hence the lightning-quick editing, the multiple superimpositions, the triptych expansion of the screen. But Tamayo also focuses on the “soft apocalypse” of the film’s treatment of faces in close-up, especially the use of the Wollensac soft-focus lens (149-56). Her analysis is superbly well-informed, incisive, and erudite. (Yes, I’m jealous.) Incidentally, I am aware that Tamayo’s work on Gance is more extensive than evidenced by her existing publications. I do hope that her research on Gance’s unrealized projects (i.e. the bulk of his creative career!) will one day be published. In me, there is at least one eager reader.
I have only one other observation about these pieces, which also applies to the 2024 book as a whole. This is the balance between new and old scholarship on Gance in the essays’ bibliographies, which are heavily skewed in favour of recent work. (And there is no general bibliography in the book.) It is as if nothing on Gance was written before the year 2000. Even Gance’s biographer, Roger Icart, gets only a passing mention. The balance between anglophone/francophone material is also noteworthy. Not counting one or two references to Brownlow’s work, I think that Tamayo’s citation of my 2015 monograph on Napoléon is the sole citation of any English-language scholarship in the entire book. These aspects of bibliographic balance speak, perhaps, to the fact that this new Table Ronde publication is not aimed at an academic market – the sources are mostly to primary, not secondary material. It is also, needless to say, aimed at a francophone market. Indeed, the book makes me wonder what kind of strategy is planned for the restoration’s international release. What kind of accompanying (i.e. written) material will be released outside France, and how will the film be released and marketed? I note that Vezyroglou is soon to publish a book on Napoléon – will Mourier also publish his own, more detailed, account of the restoration? These are questions that will only be answered later this year, when (I presume) Napoléon enters the commercial marketplace – cinemas, television, streaming, Blu-ray…
In summary, this is a very pleasing book to look through and an exceedingly interesting text to read. I regret that I have spent so much time highlighting unanswered questions about the Cinémathèque française project, but much of the film’s history is already known to me: it is precisely the unknown factors of the restoration process that interest me most! For readers who are less familiar with Gance and Napoléon, it is undoubtedly a great resource. It provides both a history of the film and a context for the new restoration. As I have tried to indicate, it still leaves some odd gaps in the information – but I must conclude by emphasizing that the restoration is surely one of the most important ever undertaken (certainly in the arena of silent cinema). I have nothing but admiration, and profound gratitude, for the monumental effort of Georges Mourier and his team. My only reservation is that the complexities of the Cinémathèque française project are inevitably simplified for the sake of commercial marketing, which does justice neither to their work nor to the film. Publicists and distributors like simple narratives, but the history of Napoléon is anything but simple.
In this context, I think the term “la Grande Version” is not particularly helpful, just as the reliance on runtimes rather than lengths confuses an already complex situation. As I have tried to indicate, the rather ambiguous discourse in the book (echoed in press releases) results in a false impression – something akin to the syllogism: “Gance envisaged a seven-hour film; the Cinémathèque française restoration is seven hours; therefore, the Cinémathèque française is the version Gance envisaged.” Mourier himself has indicated the staggering difficulties of the film’s physical and restorative history, and the work of his team in the face of these challenges is astonishing. But transparency is always the best policy, and it would be nice to see – if not in this 2024 book, but elsewhere in writing about the restoration – a more open account of some of the issues I have discussed.
Re-reading what I have written, I wonder if my reservations are only of real concern to obsessives like me? After all, I still very much enjoyed this book – and I hope the restoration generates more interest, more writing, and more publications on Gance and his masterpiece. And all my comments must be put into context: I have not yet experienced the new Napoléon in the cinema. This I will do in a few weeks’ time – and I look forward to writing more about it then…
Paul Cuff
The Realm of Silence 