Taking another look

Photographs, it has long been said, are unmoored fragments of reality, free-floating and code-less messages, ready-made to adapt to and adopt whichever meaning comes to beset them in any instance. When photographs appear as illustrations in literary texts, this “promiscuous” (Sontag, 2001: 129) referential nature, a dislocation that is inherent within them, permits photographs to be considered—much like in a collage painting—as constitutive and integral elements of a whole, rather than being read as any sort of “reductions” (Barthes, 1977: 196) of the texts that they might serve to illustrate. In intermedial studies, photo-illustrated texts can be considered, in this way, not merely as texts to which photographs have been added as complementary visuals, but rather as more collaborative or cohesive medial constellations, or

the result or the very process of combining at least two conventionally distinct media or medial forms of articulation. These two media or medial forms of articulation are each present in their own materiality and contribute to the constitution and signification of the entire product in their own specific way. (Rajewsky, 2006: 52)

Whether one applies the terms medial constellation or media product (Ibid.), synthetic intermedia (Schröter, 2012: 16) or le message mixte (Abastado, 1980: 233), the label used suggests nonetheless “[the] fusion of several media into a new [intermedium] that supposedly is more than the sum of its parts” (Schröter, 2012: 16) and is bound up in the essential premise of intermediality studies: such that media “do not exist disconnected from one another” (Ibid.: 15). Intermediality is to be understood as a “communicative-semiotic concept” that implies integration—to varying degrees—of a media product’s constitutive elements (Rajewsky, 2006: 52). It is about a medial tension in the sense that, in their copresence, “different media complete each other,” and that this tension is “[e]mbedded in intentional acts” (Oosterling, 2003: 36). Indeed, it is the suggestion of intentionality which is perhaps the most compelling argument to look toward intermediality in text-image studies: what can then be explored is how authors might engage with photographs that come to illustrate their literary works.

But photographs, too, have a temporal nature, what Barthes came to call a superimposition “of reality and of the past” (Barthes, 1981: 76). In “Rhétorique de l’image” (1964), he had first called this feature a photograph’s “having-been-there” (Ibid., 1977: 44), recognizing there always remains, with a photograph, “a kind of natural being-there of objects” (Ibid.: 45) depicted within the image. Years later, however, in La chambre claire (1980), Barthes would reconsider and refashion this feature into a photograph’s “THAT-HAS-BEEN” (Ibid., 1981: 76).1 And through this reconsideration occurs a significant ontological shift, as Barthes completely reoriented the referent to declare that a photograph’s “testimony bears not on the object but on time.” (Ibid.: 89).

Within this photographic temporality, we can see an inherent postscriptiveness in the sense that “ce qui vient après a aussi connaissance de ce qui était avant” (Jewsiewicki, 2011: 2), and that “[a] photograph is of a particular time and place, a real moment; [yet] you can make the viewer experience a slippage, a perceptual shift, so as to engage with its presentness.” (Sundarum, 2014: 338). In this essay, we analyze a photograph which illustrates the writer, ethnographer and surrealist Michel Leiris’ 1934 travel diary, L’Afrique fantôme. The photograph in question was taken in 1932 during the French ethnographic expedition, La Mission Dakar-Djibouti (1931-1933), that is documented in Leiris’ book. More recently, the photograph became the subject of an article (Novello-Paglianti, 2011) that levels a primitivist critique stemming from a legacy of 19th-century colonial ethnographic and travel photography that still held presence in the 1930s. In levelling this critique, however, the author of the article conflates the photograph’s caption and its visual content, producing a analysis of the image that, while interesting, is fundamentally flawed. We conduct a close reading of the photograph, its caption, and Leiris’ diary that the photograph illustrates, to analyze the relationship the three elements share, and which can only be maintained within the constellation of Leiris’ work. To separate these elements, as is attempted in Novello-Paglianti’s critique, ruptures their combined, constellated meaning. But Leiris’ captioning of the photographs in L’Afrique fantôme further reveals an effort to challenge the pastness of photographs, by writing them ex post facto into the subjective present of his daily journal.

La pose d’une odalisque2

Nanta Novello-Paglianti’s 2011 essay, “La construction de la représentation de l’autre. Le cas de la photographie ethnographique”, takes up ethnographic photography as its primary concern, giving close attention and individual detailed analysis to one of L’Afrique fantôme’s images. There is a solid and valuable primitivist critique in the work, a critique that stems from a legacy of 19th -century colonial ethnographic and travel photography that still held presence in the 1930s. But the approach taken here neglects the crucial importance of the text of Leiris’ journal and what it brings to bear upon the reading and understanding of the photographs. In Novello-Paglianti’s essay, the avoidance of the text of L’Afrique fantôme significantly compromises the author’s analysis of the photograph. The conclusions that result from this, we examine below, in order to highlight some of the formal and structural elements which quite inextricably connect L’Afrique fantôme’s images to the content and the context of Leiris’ journal.

In “La construction de la représentation de l’autre”, Nanta Novello-Paglianti draws attention to three photographs, two taken during La Mission Dakar-Djibouti, one of which appears in L’Afrique fantôme, and one, taken on a later Griaule-Led mission—in French Sudan (present-day Mali) in 1948. Hereby she argues that ethnographic photography functions to legitimize ethnographic discourse (Novello Paglianti, 2011: 173), and that such photography is essentially the foundation of the “expressive canon” of Western visual ethnography: “[le] regard qui sera appliqué à l’Autre pour le classifier, l’archiver et surtout l’immortaliser” (Ibid.: 174). It is a familiar argument, that of a continued tendency in Western tradition toward the construction of exotic otherness, a construction shaped by distance and by difference whose function is not to reveal any truths about itself but rather to serve as a useful, convenient, and manageable representation in the Western imagination (see Said, 1979: 60; Breckenridge and van der Veer, 1993: 6)3.

Since the Middle Ages, European explorers, in their written accounts of voyage and discovery, assumed implored that the reader believed in their words, for what they were recounting was that which they had seen with their very own eyes (Affergan, 1987: 144). Through this privileging of the visual experience, Western travel writing and, later, ethnography, pretended to an objective, observable truth. Johannes Fabian labeled this visualism the belief that one’s ability to ‘visualize’ a culture would become synonymous with understanding it (Fabian, 1983: 106). While, in both travel and ethnographic writing, visual experience has been mediated and primarily transmitted textually, the inclusion of images offered important supports to written accounts, providing a “tangible and lifelike” realism that signalled both the authenticity and the authority of the texts (Steiner, 1995: 207).

For centuries, these images were produced through hand drawings and engravings, illustrations created by “European observers who reduced [non-Western peoples] to a metaphor of Otherness that served only to confirm European expectations of the exotic rather than to challenge those assumptions” (Ibid.: 203). As Steiner notes, the illustrators were most often not the travellers themselves, but rather trained artists at home in Europe who relied on voyagers’ written descriptions along with their own artistic training to compose their images; the resulting illustrations were often wildly inaccurate composites, borrowing from and combining previous and disparate sources of exotic imagery, while applying the classical artistic styles and techniques the artists themselves were trained in, ultimately establishing new “conventional” portrayals of non-Europeans (Ibid.: 206-211). And so, for centuries, these illustrations constructed, disseminated and proliferated fictions, falsehoods and stereotypes of non-European peoples that nurtured Western primitivist mythologies.

With the advent of photography, it was presumed that technology would now allow for inarguably accurate visual proof to be provided in the form of photographic images4:

Photography was quickly recognized as a vital tool in the transmission of data, and what was thought to be reliable data at that. Photography’s chemical connection to what it depicted, the fact that, as Benjamin wrote, it was ‘seared with reality’5, suggested that it might be capable of capturing and conveying ‘facts about which there is no question’6. While the anthropologist was sundered from the ‘man on the spot,’ photography was recognized as a crucial mediator. (original footnotes) (Pinney, 2012: 15)

However, it is widely argued that the same fictions and falsehoods, the same metaphors of Otherness persisted all the same. Childs writes that through the latter part of the 19th century, photography became “a crucial element of control and visual privilege,” and that even beyond being a convenient and affordable means of communication, the photographic image became “an ethnographic source that authenticated the appearance and culture of the Other in the colonial sphere” (Childs, 2001: 52). The newly ’photographed’ Other would thus be integrated into the workings of the colonial apparatus, promoting the “colonial idea” (Chafer and Shackur, 2001: 1) at home in travel books, in magazines and the press, and in the general consumer and tourist marketplaces, through collectibles like postcards and photographic prints of ‘exotic’ peoples and places.7

In “La construction de l’autre”, Novello-Paglianti criticizes the othering powers of this ethnographic and photographic gaze, focusing specifically on the use and representation of the “native” body as evidence of scientific proof and legitimacy: “[L]e corps doit être une donnée objective,” she writes and even adds: “Les corps des populations sont représentés pour leur ‘enveloppe corporelle’ perçus pour leur surface. La simple monstration du corps fait épreuve de diversité” (Novello-Paglianti, 2011: 177). It is this concern that governs her selection of images—all three featuring native bodies as their subjects—to argue how the ethnographic photograph is pure construct, how it perpetuates the Western canonical image of otherness. In her analysis of the photograph that she selected from L’Afrique fantôme,8 she states that

la photo intitulée “Jeunes gens filant le coton en courtisant les femmes (Poli, 25 janvier)”9 [Leiris, 2014: 253] remet en scène deux thématiques chères à l’ethnographie. Il s’agit de la représentation des autochtones qui sont en train de filer le coton. Au premier plan, on note un homme assis par terre occupé dans cette action de filage avec des moyens traditionnels: une main qui teint la matasse de coton et l’autre qui tend le fil. Au deuxième plan, un autre homme, à demi-couvert par le premier, a les mains occupées dans différentes actions et, au fond, nous apercevons une femme dans une position oisive. Elle est semi-allongée et elle est la seule figure à ne pas être impliquée dans quelque occupation. Notre attention est centrée sur le premier homme, donc les bords de l’image, qui sont marquées par une autre figure masculine assise dont on [ne] peut entrevoir que la moitié du corps, n’attirent pas notre regard. L’évocation de la séduction se fait à travers la seule figure féminine présente qui incarne plus la pose d’une odalisque que la courtisanerie masculine.D’une part, nous notons la division entre eux et nous, eux photographiés là-bas, et nous “simples” spectateurs et chroniqueurs de mœurs. D’autre part, on assiste à une réévaluation de l’idée de l’authentique qui devrait être, pour l’époque, un des traits de la vie “sauvage” et la volonté de l’anthropologue de ramener dans son pays une partie de cette authenticité. Le sauvage représenterait une diversité qui garderait [en] son intérieur une idée de pureté.Le mouvement cueilli dans ces clichés d’autochtones affiche des corps qui agissent et qui ressentent, qui éprouvent des émotions, etc. Ce vécu a ici valeur de vérité, de témoignage réel. (original emphasis) (Novello-Paglianti, 2011: 180)

That Novello-Paglianti has selected this image over others is understandable. As she argues, the figures are in motion, the “bodies” here are acting, being, living. And for this, she is at once admiring and critical of the photograph: she acknowledges the captured moment’s “purity” while admonishing the ethnographer/photographer for mythologizing that very purity as yet another “savage” or primitive trait (Ibid.).

However, beyond the image’s subjects—these active and emotive bodies—, the photograph’s rather high aesthetic qualities further contribute to the constructed reality that Novello-Paglianti challenges. The image, titled “Jeunes gens filant le coton en courtisant les femmes” is quite possibly the most visually compelling image published in L’Afrique fantôme: its composition is strikingly well-balanced and engaging—a unique feature, we would suggest, among the thirty-one images that illustrate the journal—; the architectural lines of the beams supporting the pergola under which the young men are gathered are repeated near-perfectly in the foregrounded figure’s upwardly stretched arm, and their angle is repeated again in the thread of cotton that the figure pulls and manoeuvres in his hands. The camera’s focus is soft, the light is hazy and diffuse—the image itself is charming, seductive. Even the photograph’s caption, “Jeunes gens filant le coton en courtisant les femmes”, is noticeably poetic, having internal rhymes and alliterations that point to Leiris’ penchant and talent for wordplay.10

In fact, we would argue that the caption11 is actually the most significant element in play here. Earlier in her essay, Novello-Paglianti wrote that

la légende qui accompagne les photographies est essentielle pour [l’authentification]. Chaque assertion figurant à côté de l’image est acceptée comme vraie. Le discours verbal a le rôle de légitimer la trace indicielle de la photographie. Une des caractéristiques de la photographie ethnographique est que le verbal ne doit pas rentrer en opposition avec l’iconique. Le discours verbal semble surgir de celui iconique qui en même temps le confirme grâce à son caractère mimétique. Le mécanisme se construit comme une boucle dans laquelle chaque partie soutient l’autre grâce à un principe de compatibilité recherchée. […] L’anthropologue détient ici tout son savoir et son autorité d’explication. […] Enfin, le lecteur a besoin d’être guidé dans son interprétation par un expert qui peut lui expliciter les moeurs locales (Novello-Paglianti, 2011: 178).

And while the L’Afrique fantôme photograph was selected by Novello-Paglianti for its display of the human form, it is the caption, more than any other element, that guides her analysis and interpretation of the image. However, this reading that she ultimately extrapolates from the photograph comes from a rather significant misrepresentation of the scene.

Reading the caption, “Jeunes gens filant le coton en courtisant les femmes”, one is informed of two activities—the spinning of cotton and the courting of women—, which, the caption suggests, are being performed in tandem by the young men in the photograph. The first of these activities is readily apparent in the image: one sees an outstretched arm, its hand grasping a white thread of cotton, the line of which draws down the centre of the image to connect with another busy, motion-blurred hand. In this respect, the relationship between image and caption is exactly as Novello-Paglianti argues it should be: the caption, conjured from the photograph, mimics and confirms the visual display.

The second activity, courtship, which may often be more social than demonstrably physical, proves to be more elusive. In the photograph, one sees no overt displays—nor even hints—of flirtation, attraction or seduction. In fact, one catches perhaps the very opposite. The first two figures appear fully and simply engaged in their work: each faces forward, back turned to the figure behind, eyes focused upward or downward upon one’s own busied hands. The third figure, conversely,seems to be rather disengaged from any activity: body reclining, face expressionless, gaze adrift. In this respect, caption and image appear to be fully at odds. Yet, in her analysis, Novello-Paglianti wishes, nonetheless, to maintain the relationship between caption and image, but to do so, she is required to resituate the visually-absent courtship away from the two young men busily spinning cotton, placing it instead upon the third, idle, reclining figure.

Framing this third figure as an “odalisque”—described by Novello-Paglianti in this way due to the languid pose, reminiscent of European orientalist paintings such as Ingres’ La Grande Odalisque (1814) and L’odalisque à l’esclave (1839), paintings in which “a nominally eastern woman lies on her side on display for the spectator” (DelPlato, 2002: 9)— Novello-Paglianti suggests the photograph, then, to be more an image of subjugation than of courtship. Essentially, what she is offering here is a criticism of the French mission civilisatrice and of the general machinations of colonialism, which, in itself, is certainly fair and warranted—the main thrust of her essay, after all, is an examination of the ethnographer’s exercise of colonial power (Novello-Paglianti, 2011: 183)—. And so, it is not surprising to find, within the article, criticism suggesting the image to be yet another in the continuing stream of colonial-era stereotypes of non-Western peoples (see Steiner, 1995; Childs, 2001).

However, to be able to arrive at this conclusion, a number of elements have been misconstrued. Firstly, the supposed submissive nature of the photograph is one that Novello-Paglianti herself has constructed. In her reading of the image, she had to alter the active meaning of the verb “courtisant” in the caption to have it represent rather more passive qualities that are better suited to the related noun courtisane. The impacts of this transposition are significant: whereas the caption attributes all agency to the young men in the image—they are spinning cotton while they court the women—, Novello-Paglianti’s reading revokes this agency, giving it instead, and exclusively, to the photographer/spectator in the form of a “colonial gaze” (Alloula, 1987: 92) that is being cast, desirably, upon a submissive, feminized Other (see Said, 1979: 137-138).

Secondly, the “odalisque” figure that Novello-Paglianti centres upon is also quite clearly a man—the women who, the caption suggests, are being courted, do not even appear. Granted, the odalisque figure’s reclining pose and tunic-covered torso—in contrast with the other two more active and exposed male figures—can be read as feminine, but such a reading is derived from a Western-coded image of femininity (e.g. from Ingres’ paintings) and does not necessarily reflect the types of gender performance expressed by the Kirdi people who are the subjects of the image. Indeed, in the text of L’Afrique fantôme, Leiris explained about the Kirdi that “[l]es femmes sont nues, à l’exception de deux touffes de feuillage”12 (Leiris, 2014: 252 [24 janvier 1932]). This simple statement, which is supported by a number of photographs taken by the Mission though not published in Leiris’ book,13 suggests that Kirdi women (at least those encountered by Leiris) tended in their dress to remain exposed above the waist. It is also telling of Novello-Paglianti’s own reading of the photograph, showing, in part, that her interpretation is the result of having isolated both image and caption away from the text of L’Afrique fantôme.

Leiris’ description of the Kirdi women’s dress appears in the entry for January 24, 1932, which was just one day before the photograph was made, and it refers to the same village, Poli, and community of people that feature in the photograph. The description does not, by itself, refute Novello-Paglianti’s reading of the photograph, but its close proximity in the journal and its rather direct relationship to the subjects of that photograph immediately raise questions as to the degree to which the text may have been consulted in informing Novello-Paglianti’s analysis. And these questions only become more pronounced when one reads the entry for the day on which the image was actually captured. Writing this entry dated January 25, 1932, Leiris begins, “Nouveau tour au village kirdi. Sous un abri de tiges de mil, des jeunes gens filent le coton à côté des femmes. Peut-être leur font-ils la cour. Nous offrons une tournée de pipi.14 Tout le monde y participe […].”15 (Leiris, 2014: 252) In this opening passage, one reads the line from which the photograph’s caption is derived: “Sous un abri de tiges de mil, des jeunes gens filent le coton à côté des femmes. Peut-être leur font-ils la cour.” What is notable is that this is merely a passing observation, a suggestion—maybe they are courting the women—, one that is made quickly, casually, before moving on to other details and recollections from the day, which will include a shared round of the alcoholic drink “pipi”, then on to Leiris’ complaining that travel by horseback is painful and, therefore archaic, and Leiris being shown by a blacksmith, in mime, the Kirdi manner of circumcising a penis: “ainsi qu’on pèle une banane”16] (Leiris, 2014: 253).17 This day’s entry, like the vast majority of the diary entries in L’Afrique fantôme, is personal, filled with the mix of wit and cynicism that, throughout the journal, tends to be Leiris’ stock-in-trade.

That the photograph’s caption is derived directly from this entry and not conjured from the photographic image is significant in a number of ways, as it is ultimately this fact—this relationship—that challenges the tenability of Novello-Paglianti’s reading. Being a reiteration of the diary, the caption is, more than anything, an extension of the text that cannot simply be read as a mimetic description of the photograph. Nor, in this sense, can caption and image be extracted away from the text, both of which Novello-Paglianti had done.

To borrow Novello- Paglianti’s own model for a moment, it is not two, but indeed these three elements—image, caption and text—that constitute the loop of mutual confirmation, the mechanism that, in her argument, codifies or legitimizes the photograph (Novello-Paglianti, 2011: 178). And with three as opposed to just two elements in play, the apparent ambiguities between the photograph and its caption, these ambiguities that forced Novello-Paglianti into a variety of semantic acrobatics in order for them to be reconciled, are dispersed.

Writing captions in and as postscript

Simply speaking, the text of L’Afrique fantôme offers a much bigger picture with which to inform, contextualize, or situate the book’s photographs, as the photographs are understood here to be read, not as complete or separate documents unto themselves, but as elements of the larger whole. In the case of the photograph “Jeunes gens filant le coton en courtisant les femmes”, while the women mentioned in the caption remain absent from the photograph, they are fully present in the text. Furthermore, the text having situated the women as being “à côté” in relation to the young men, they are easily, and unproblematically, situatable outside the frame of the image without being absented from the scene18.

Similarly, with the courtship mentioned in the caption being quite underplayed in the text, the impulse to seek it out visually in the photograph is also lessened, perhaps even quelled entirely. And all the while, the photograph continues to perform its function of visually confirming the presence—in both caption and text—of the spinning and courting young men; and with mutual confirmation being equally returned from both—and not just caption alone—, their agency, too, remains securely intact.

There are some instances in L’Afrique fantôme where a caption will provide information which is neither available in nor expressly relevant to the visual element, such as in the caption that reads “Un coin du quartier Saint-Michel où habitait la borgne Dinqié”19 (Leiris, 2014: 577). As the photograph in question is a wide panoramic shot taken from above that includes separate clusters of ruins across the foreground as well as a spate of thatched roofs that extends in a line toward and then across the horizon, one is surely not able to identify which anonymous dwelling may belong to the woman named. But, having read the journal and arriving at this photograph, one will have become quite familiar with the one-eyed woman Dinqué, and have also been made privy to some of the melodramas that have unfolded across the city, whose skyline is pictured, and in Dinqué’s very neighbourhood, including a fight among schoolchildren which was instigated by Leiris himself and an attempted murder which was possibly the result of a love triangle that had soured (Ibid.: 578). The caption, here, favors the conceptualization of the image rather than the simple identification of its contents; we find, again, thatthe caption is not offering a mimetic description, but far richer memories of place.

In another instance, Leiris interjects even more directly into a photograph’s caption: “Visite du guérisseur Alaqa Taggagn (l’homme au parasol) à Malkam Ayyahou (tête couverte, debout devant le seuil de sa case hôpital). La scène—qui se passe dans le cours—est vue du perron où je couchais (13 octobre)”20 (Leiris, 2014: 556). What is actually surprising about this caption is that it does begin very much as mimetic description, complete with added precision to identify individual figures pictured in a crowd of several. But the intrusion by Leiris himself into the caption—“du perron où je couchais”—changes the operation of the caption entirely. It draws the reader/spectator’s focus away from the crowd which is central to the picture and positions them, as Leiris, onto the porch, apart and excluded from the scene playing out before the camera.

Despite the fact that it is ‘behind the camera’ where the reader/spectator essentially rests in regard to a photograph in all situations, we would argue it is rare for a caption to actively and intentionally work on maintaining that distance. Rarer still is a caption, such as this one, that first draws a reader/spectator into a scene only to pull them back to its periphery. But this caption, too, is functioning more in service of the diary than of the photograph. The distance the caption creates for the reader/spectator actually mirrors the very real exclusion that Leiris experienced—and was slightly wounded by it—on the day in question, as his diary notes:

Bien que couché encore à 1 heure du matin passée, je me suis levé de bonne heure, désireux d’assister à la cérémonie promise. Mais, arrivant à la maison des wadadja, j’apprends qu’elle a eu lieu sans moi. Nous étions encore endormis et l’on n’a pas osé nous réveiller. Il n’y avait du reste pas grand-chose à voir. […] Pour pallier ma déception, Malkam Ayyahou recommence la cérémonie pour l’herbe qui est restée sur les banquettes. Elle sort en tête de la procession et revient en queue. Danse guerrière en rentrant [13-10-32].21 (Ibid.: 572-573)

Whether they are direct extensions of the text, as with “Jeunes gens filant le coton…”, or they imprint Leiris’ full subjectivity onto the photograph, as with the intrusive “où je me couchais”, what these captions are doing, essentially, is writing the images, postscriptively, into the journal’s narrative to make them as personal and subjective as the diary itself. Therefore, the captions operate to challenge the inarguable and inevitable pastness of the photographs in their attempts to have the reader/spectator “engage with [the photographs’] presentness” (Sundarum, 2014: 338) in the same way the reader engages with the diary’s presentness, through its steady and daily real-time advance across a continent over months and into years.

The captions in L’Afrique fantôme are less involved in the “veridiction” (Baptista, 2006: 261) of the photographs, which is the truth-affirming assumption under which Novello-Paglianti conducted her analysis, than they are in establishing authorship (see Scott, 1999: 82) over them. And what is meant by authorship, here, is not photographic credit22. Rather, authorship, in this sense, is about authority, about control over the narrative of the whole work. Actual credit for the images published in L’Afrique fantôme rests with the Mission’s leader, Griaule23. And by this attribution, the photographs are official, sanctioned—like the Mission itself—; and it is in this way that they register as ethnographic documents. It is as an ethnographic document that Novello-Paglianti attempts to address the image captioned “Jeunes gens filant le coton en courtisant les femmes”. The caption, however, resists this status and complicates her analysis. The casual partialness of its observation, the form and structure that are more poetic than scientific, and the indebtedness it has to the text all demonstrate how the captions in L’Afrique fantôme actively operate to appropriate the photographs away from this official attribution, away from a documentary or ethnographic status (Rushton, 2012: 170).

Novello-Paglianti’s reading of one of L’Afrique fantôme’s photographs is compromised not because of a theoretical error, but it fails to recognize or to consider that the most important mediating actor upon the photographs is the journal itself. And the caption, here, owing more to the diary’s description of events than to the image contents, demonstrates perhaps most effectively the extent to which the text of L’Afrique fantôme operates as the “dominant informative element” (Baptista, 2006: 259). Not every caption in the book weaves together image and text quite as poetically as “Jeunes gens filant le coton en courtisant les femmes”, but we would argue that, by and large, they hew more closely to the content of the text than to the descriptions of the images they support.