Antinomies of representation

This work is licensed under the Creative Commons | © David Zeitlyn. ISSN 2049-1115 (Online). DOI: http://dx.doi.org/10.14318/hau4.3.022


Antinomies of representation

Anthropology as an ekphrastic process

In memoriam: Rigobert Blaise Tueche August 26, 1968–December 8, 2000

David ZEITLYN, University of Oxford

This article addresses a profound anthropological issue: how do representation and the represented relate? What motivates or warrants the inevitable disconnection? It is a mistake to dismiss representation as misguided, oppressive, or misleading. Representation is part of cognition generally and natural language in particular. As such it is inescapable and part of how we think and talk about the world. Moving between visual and linguistic anthropology I suggest that photographs and portraits provide a rich basis for thinking about the particular sorts of warrants for anthropological representations. The general conclusion is that anthropological representation may be conceived of as a form of ekphrasis (a verbal account or evocation of a typically non-present image or object) providing the indexical or deictic bridge between representation and the object represented. As “similarity implies difference” so “representation implies ekphrasis.”

Keywords: ekphrasis, representation, photography, visual anthropology, linguistic anthropology, Cameroon

Description is revelation. It is not
The thing described, nor false facsimile

It is an artificial thing that exists,
In its own seeming, plainly visible,
Yet not too closely the double of our lives,
Intenser than any actual life could be

—Wallace Stevens (1997)


In visual and linguistic anthropology, representation is construed in very different ways. In this paper I discuss and employ some of the theoretical ideas culled from both traditions as I consider research material produced from ongoing field research with Cameroonian studio photographers.1 Reflecting on the general topic of social science representations (accounts and generalized statements about social groups and phenomena), I consider ways in which a deictic connection between representation and the object represented may be conceptualized.

The epigraph above frames the discussion which follows. My suggestion is that, in Wallace Stevens’ terms, anthropology provides warranted facsimile rather than revelation or empathic identification, and that photographs and portraits provide a rich basis for thinking about the particular sorts of warrants for anthropological facsimiles. The general conclusion is that anthropological representation may be conceived of as a form of ekphrasis (a verbal account or evocation of a typically non-present image or object; I discuss the term below). This helps us think about the mysterious deictic (indexical) relationship between representation and the object represented.

The discussion takes place in the context of the concerns about representation that have long been a source of existential angst in anthropology and the social sciences. Having considered how visual material exemplifies and magnifies the question of what it is to represent someone or some social group, I take some ideas from ethnomethodology (repairs), sociolinguistics (deixis), and art history (ekphrasis) and use them to re-examine the idea of anthropological representation. I take portraiture as a critical case in point since portraits are exemplary representations (of course, they may be other things as well, see Zeitlyn 2010a). I consider the treatment of images after death in Cameroon, the display at interments and post-mortem celebrations and the annotation of existing photographs by friends and family to mark an image as containing a dead person. To connect back to the general topic I reveal a parallel between the Church of Latterday Saints retrospectively baptizing people into their church and the effects of putting names to images, adding captions: opening images up to new futures, new uses.

Representation always has a deictic or indexical element, which bridges the representation itself and the represented (see Silverstein 2003). This can be thought of in terms of the change that occurs when identifying a person in a photograph: from an anonymous image to a portrait of a known individual. So social science representations are ekphrastic in that they are translations across media; they are also indexical: in anthropology, pointing from the multiple heterogeneous results of fieldwork to the polished prose of an anthropological account; in Stevens’ terms it is revelatory, providing warranted not false facsimile. Deixis is the general term for indexical uses of language, either of place (e.g. here), time (e.g. now), or person (e.g. me). It is of great interest to social linguists because it is an excellent example of how meaning (or at least, reference) is achieved by usage: the referent is determined by the place, time, and person speaking so that even grammarians have to consider the context of usage. In another context (2005b), I have explored social deixis (the labelling and identification of social persons through pronouns, names, and kin terms). There are similarities between the linguistic phenomena of deixis and how a description connects to the image or thing described. In deixis an indexical term achieves its referentiality in performance: the pronoun “me” in this utterance points at me, the speaker, but when another person says “me” it refers to someone else entirely.

The ekphrastic process (the representational process) is similarly indexical or deictic; it is a social achievement, the realization of a commonplace but critical task: connecting words and images, descriptions with the things described. To see how the process works we need to look beyond and behind the descriptions themselves, to lay bare the mechanism by which a description becomes connected to its object, and how it maintains that connection. We must consider representation, not just philosophically and linguistically, but also at a social level. We need to look at how a description is taken to be such, at what Harvey Sacks (1995 vol. 2: 3–16, 249–68) called the “second story”—i.e., the account which justifies and establishes the deictic process. As Bryan Wolf (1990: 189) puts it, “The issue, then, is representation—the manner in which objects, visual or verbal, present themselves. And the question that accompanies it is: by what rules does representation operate in the modern world?”

Introducing ekphrasis

Many of these issues hark back to classical debates about the very idea of representation. The key Aristotelian distinctions are between display or showing (mimesis), recounting or narrating (diegesis), and explanation (exegesis). To that triangle of Greek terms I add a fourth: ekphrasis. Ekphrasis began as a term in rhetoric for the description of nonexistent works of art (either mythic objects such as Achilles’ shield in the Iliad, or of lost classical works of sculpture2). It was “originally used by Greek classical rhetoricians to qualify a description with great visual content” (Ramos 2004: 147). From this it has been generalized to include the use of language to describe images or to evoke imagery. In the philosophy of science this is fundamental. For Ian Hacking (1983: 145): “In physics and much other interesting conversation we … make representations—pictures in words.” He argues that the base definition of what it is to be human is homo depictor, in which the making of likenesses is fundamental. This stance is provocative, particularly when applied to the problem of representation. However, there are further problems. First, Carrier (1987) distinguishes ekphrasis and interpretation. On his account, ekphrasis is descriptive and/or evocative, and interpretation analytical. As a consequence, Carrier suggests, we can only really quarrel with interpretations. This is another version of Worth’s (1981) “Pictures can't say ain't” argument: that images cannot make logical propositions such as negatives and hence cannot construct arguments in a strict sense. The claim is that one cannot express the statement “I disagree” visually.3 This point is captured by the classical contrast between diegesis (as description) and exegesis (explanation on the basis of understanding).4 Before continuing with the theoretical discussion we should consider a concrete example of representation, a photograph taken by a Cameroonian photographer for local clients.

Figure 1
Figure 1. Chief Mɔgɔ Michel (d. 1977) and retainers c. 1970. Annotated by the late Gamgbe Lucas who is shown, standing back right. Used with permission.

An example from the field: Marking prints

The image above illustrates some of the issues about representation that I am concerned with. It was taken by an itinerant photographer on a market-day in Somié village around 1970 (unlike many such prints it is unmarked with either date or photographer’s stamp). Chief Mɔgɔ is shown with his close attendants, a wife (the second woman is unknown) and one of his brothers (who had retained a print and annotated it, then later discussed it with me).5 Originally there were probably five or six prints but only one seems to have survived into the twenty-first century.

The image is typical of the output of Cameroon’s black and white photographers who were active until the widespread adoption of color in the late 1980s. Their main activity was taking passport style photographs, which were needed for ID cards, school cards, and many other official documents. These requirements had the effect of securing the economic basis of photography and kept the prices down for identity card photos and for other styles as well. Many other types of photographs were also taken, in order to mark births, deaths, marriages, the acquisition of prestigious items such as radios, motorcycles, or sewing machines, or to mark the visit of a family member from another town or village, such as the image above.

We could give many accounts of this image. These would concentrate variously on 1) its composition as a collaboration between the people depicted, the photographer and the (presumed) onlookers; 2) its life history since being printed—moving with its owner from the village to Tibati some 300 km away for many years before returning to Somié; 3) the socioeconomic milieu in which the photographer was operating—he almost certainly visited on market days, taking the image one market day and returning with the print the following (although some photographers would print on the spot using the box camera as printer/enlarger); or 4) the problems I encountered during fieldwork locating images of Chief Mɔgɔ(see Zeitlyn 2015 forthcoming). There is another aspect that I wish to focus on here: the annotations, one in red biro on Tchimbi’s chest (standing back left) and the other on Chief Mɔgɔ’s chest in pale blue biro. Such annotations are common in Cameroon: I have seen them repeatedly on many images across the southern half of the country. The practice is clearly widespread and not restricted to a particular cultural tradition (or even a country: it is also found in Ghana, see Bajorek 2013: 67).

Below I discuss some general aspects of images of the dead. What motivates that discussion is my continuing perplexity about this practice of annotation6 (which I suspect I have never been able to convey clearly to my colleagues in the field either speaking Mambila, French, or English). It clearly feels properly respectful to a senior relative or a personal friend to “mark their passing” by marking their image. But when looking at a photograph, the person we see is and is not changed by death. Perhaps the ambiguity is reflected in Kendall Walton’s (1984, 1986) argument about whether we look through photographs as through spectacles (or use mirrors to see round corners), so we really are seeing the people depicted, across time and space (but note Martin’s 1986 disagreement). We don't need to agree with Walton in order to appreciate the point about representation that he is making—i.e., that some mediated perception is commonly taken as direct perception for all its mediated qualities. Few people would disagree with the claim that they see themselves when looking in a mirror. So an image of someone we know to be dead is annotated respectfully as a form of action at a distance, in the way we might otherwise throw a handful of soil into their tomb at a burial, or, in the Jewish tradition, leave a stone on their gravestone. This is a form of sympathetic action much analyzed throughout a century or more of anthropological theorizing. Without wishing to get embroiled in an argument about the status of sympathetic action and its relation to sympathetic magic, I think it is clear that these debates have some relevance here. As I said above, a photograph both is and is not altered by death. If I can no longer see someone in the flesh then my only option is to gaze at pictures, still or moving, and let those be emotionally moving precisely because I cannot compare them with the original. Yet at the same time nothing about the image has changed and perhaps that is the discrepancy which motivates the annotation: we know that a person has changed (died) and so we alter the image to maintain the ongoing relationship of surrogacy between image and its object. Could we, perhaps, view this as restoring the facsimile from falsehood, as in Stevens’ poem?

Some metatheoretical background

Discussion of surrogacy, facsimile, and the relationship between image and object are ways of addressing various problems of representation, which have been separately discussed in philosophy, art history or visual studies, and in the philosophy of science. Indeed, problems of representation dog the social sciences. Social science accounts and explanations are couched in terms such as class, ethnicity, and economic status, which are our representations of how the world is. Part of the epistemological and methodological problem is that we only see individual tokens (such as “a cat”) and not types (the generic “cat”). On what basis do we use generic terms? This conundrum affects even those undertaking large-scale questionnaire research (Williams 2003). The approach to such problems of representation affects the way that research is conducted and the kinds of record kept as “data.”

We urgently need new theoretical paradigms of what constitutes representation in order to ground the discipline somewhere between the extremes of an “anything goes” relativism and simple minded realism (often castigated as “positivism” without reference to or appreciation of Comte’s arguments). Postmodern approaches to this are increasingly seen as being as limited and as flawed as those of the old school that they sought to replace. As Lakoff and Johnson (1999: 233) put it, “eliminating simple minded realism does not eliminate all forms of realism, and it does not require either idealism or total relativism.”7 Since visual anthropology is centrally concerned with nonlinguistic representation it poses a set of particular challenges to problems of representation in the social sciences, for which the idea of ekphrastic social science may help resolve (see discussion below). These challenges mean that the concept of representation must be theorized as not a purely linguistic issue.8

When we write an account of X (for example, the “culture” of overwork among UK administrators, or attitudes to illness among a specific social group) from a social science perspective, what are we doing? There are many possible responses. Five such, chosen to illustrate the breadth of the range, include:

  1. drawing a picture (to use the visual metaphor criticised by Rorty, among others);
  2. giving an autobiographical account of my involvement, my subjective opinions, or presenting my understanding (situational and biased as it may be);
  3. identifying and modeling the underlying “basic” structures;
  4. describing how the subjects understand the events;
  5. acting with and on behalf of the people concerned in their dealings with the wider world.

It is important to recognize that such approaches are not necessarily mutually exclusive (some, but not all of the above, are consistent with one another), and to resist the temptation to advocate for a favored position to the exclusion of all others. Elsewhere Zeitlyn (2009a) has argued for a recognition of incompleteness and partial views, where an account may be complemented by others which are equally faithful to the subject at hand, and may be equally incomplete. Bruno Latour (2002), when discussing varieties of iconoclasm, conflicting images, and conflicting attitudes to images, similarly tries to avoid positions suggesting that imageless representation is possible, and thus argues against “freeze-framing.” He wants us to accept our position within cascades of images in which there is no possibility of stopping the flow and standing outside the chain of images, as it were in a relation of pure apperception, directly connected to what is being represented. His example is astronomy in which without theory-generated technology (telescopes of one sort or another) and the imagery they produce, one cannot understand the galaxy. There is no “view from nowhere,” no ideal or nonimagistic representation of the Milky Way. Appreciating this is to recognize a form of ekphrastic relationship in all representations. Accordingly, just as “similarity implies difference” I am suggesting “representation implies ekphrasis.” Ekphrasis is what takes us across the indexical bridge, as it were, from representation to the represented. My emphasis would be on what a representation does as part of human interaction (as part of social interaction) so my philosophical affiliation is more with the late Wittgenstein (1953) than with Quine (1960).

When indexicality fails

As an anthropologist I delight in thinking across the grain, in trying to examine that which is usually taken for granted and which is therefore unexamined and undiscussed. It is in moments of failure that we know ourselves best; when things go wrong, our attempts to remedy them often illuminate how we believe things should be more clearly than anything that happens when things are as they should be. When I say something and my audience does not understand me then I must repair things; I have to ensure that the connection is made (or remade) between the description or account and the thing described or accounted for. The connection is indexical, just as names are: there are fewer names than people and processes analogous to baptism establish the connection between name and the named (an idea which Kripke [1980] placed at the heart of his account of naming). Such recognition of the importance of repairs is one of the key contributions of conversation analysis, and those working in the wider field of ethnomethodology would say, along with Holmes and Marcus (2005), that it is the actors rather than the academic analysts that are “doing the sociology.”

What is particularly helpful about the way problems are repaired is that they often result in implicit understandings being made explicit, which is a boon for the analyst: the actors are “doing the analysis.” It is at this point that the traditions of the ethnography of communication meet that of conversation analysis (ethnomethodology). There are constraints about who can say what in different circumstances and to different audiences. As anthropologists are wont to point out to philosophers and linguists that speech is used in many different ways all the time, and we should not privilege one sort (giving neutral, accurate descriptions) and disparage or ignore the other sorts. There are reasons why such idealized forms may be important but this needs to be demonstrated rather than being assumed. One concrete example of this is the way in which speakers design their utterances with regard to Grice’s Maxims for linguistic cooperation and communicative implicature (1975). The peculiarity about the Maxims is that they are universal yet systematically violated. This is the starting point of Brown and Levinson’s (1978: 100) theory of politeness; they take Grice’s Maxims

to define … the basic set of assumptions underlying every talk exchange. But this does not imply that utterances … must meet these conditions, as critics of Grice have sometimes thought. Indeed, the majority of natural conversations do not proceed in such a brusque fashion at all… . Politeness is then a major source of deviation from such rational efficiency, and is communicated precisely by that deviation. But even in such departures from the Maxims, they remain in operation at a deeper level. It is only because they are still assumed to be in operation that addressees are forced to do the inferential work that establishes the underlying message and the (polite or other) source of the departure— in short, to find an implicature, i.e. an inference generated by precisely this assumption… . There is a basic assumption in talk that there is underlying method in the madness.

This may seem to be taking us too far from the practice of doing social science, but I see it as deeply connected in the unspoken assumptions we make which rarely come to the fore. So when I say “this is a description of an old photograph,” or “this is an account of Mambila witchcraft accusations,” my audience usually has no problem about what sort of activity is going on (giving a social science account), but there could be questions. There could be concern about the very activity of “giving descriptions” since, despite its taken-for-granted everyday-ness, there is a space or distance between account and object and the relationship which connects the two can go awry. Again we should recognize that Garfinkel’s ethnomethodology stressed the silent assumptions surrounding mundane social interaction. His “breaching experiments” in which students purported not to understand basic social conventions are a case in point. The anger and incomprehension which greeted attempts to answer fully questions such as “How are you?” (see Garfinkel [1984] and discussion in Zeitlyn [2004]) reveal how normal coconversationalists mutually understand that a greeting question is not a sincere request for information about one’s state of health. There are culturally specific understandings, usually unstated, about how language is used in different social contexts.

Such points are relevant to the topic of description and representation. Descriptions may go awry in many ways and one such type of problem is when indexicality fails (when it is not clear what is the actual object of description). Beyond the influence of academic philosophy, descriptions are usually taken to have relatively unproblematic relationships to their objects.9 Indeed, the assumption that descriptions can be connected to their objects is the basis for criticizing a description as “inaccurate.” This underpins the anticolonial critique of Asad (1973) through Said (1978) to Taussig (1980 and 1993) where orientalist accounts are bad because they are wrong and misleading. These authors all imply that a better, more accurate and politically anticolonialist descriptions are possible. Engagement includes engaging with forms of realist language, in other words, with description.10

As Roger Just and I have argued elsewhere there are many forms of middling realism between full blown relativism and naïve realism (see Zeitlyn and Just 2014). However, a choice must be made either to describe or to analyze the possibility of description. It is a practical impossibility to do both at the same time, so metadescription is often an academic luxury which life-in-the-world cannot afford while engaged in living it.11 Having accidentally locked the cat in the house as I leave the country, I need to tell my neighbor by telephone where to find the cat food. A verbal description must enable her to enter the house, disable the alarm, and locate biscuits and the cat’s bowl. The style of description is transformed by mobile telephony. What and how I tell her depends on whether I am talking to her on a fixed line, perhaps hours in advance when she’s in the office, or whether I am speaking to her on her mobile telephone as she faces my back door. In Garfinkel’s terminology, the description must be adequate for the task at hand. The description of how to enter the house is an example of spatial deixis in which, as William Hanks (1990) has explored, speakers deploy a variety of frames to achieve coordination between speakers and actions in the world, depending on the social context of the speech event, which includes the different social statuses of the speakers. Of course, in other social contexts, other types of linguistic usage prevail. I am by no means assuming that the giving of accurate descriptions is typical or normal. That said, I would suggest that this is an explicit goal of much social science practice.

What is generally not discussed is what David Reason (p.c. building on Sacks’ [1995 vol. 2: 3–16, 249–68] “second story”) calls the “morality of description”— i.e., the reason why the description is being produced. This involves asking for the purposes of and justification for the description (in my example above, the hungry cat). Usually there is a very good reason for this silence and lack of reflection: descriptions have goals and it is important to achieve them. The goal is usually shared by the descriptor and those receiving the description so it does not become a topic of discussion; this only happens when the description “fails to work” and repairs become necessary. Garfinkel’s breaching experiments rest on apparently not sharing communicative goals. To be clear, I am not assuming that the only or even the main purpose of general language use is to communicate descriptions. Much of the ethnography of language makes that very clear indeed. However, in this essay I am reflecting on how anthropologists use descriptions in ethnography, and the usually unspoken assumptions about indexicality in ethnographic writing, for it remains undertheorized even with all the discussion that followed the publication of Writing culture (Marcus and Clifford 1981).

However, we might also note that when Roland Barthes ([1977] 1981) discusses the unsettling detail or viewpoint which disturbs the viewer of a photograph (he terms this the punctum, that which pricks), he is seeking to achieve a different task. He wants to explain why an aspect of the image disturbs the ordinary reading, and disturbs not just him but also other viewers. For all the subjective viscerality of a punctum, it has an intersubjective force. It is not idiosyncratic to one viewer alone, so viewers can recognize the punctum of a shared image. However, when that image is not available to the viewers then the task becomes a rhetorical one, as is the case of Barthes’ (ibid.) discussion of a photograph of his mother as a child in a garden. His account is ekphrastic in the classical sense. It is rhetorical since he refuses to show it to the readers, for all the prose he expends describing it. In such a case we can never know if the deixis (see below) is achieved. This makes it hard to characterize the description. Has indexicality failed? Was it ever possible?12

Ekphrasis again

Returning to the idea of ekphrasis, we now we must consider what Bryan Wolf (1990), in a wonderfully written and deceptively anarchic article, calls “closet ekphrasis.” This discusses connections between realism and display. For Wolf closet ekphrasis concerns the silencing of art; the rhetorical claim that the visual is silent, because it is outside any language. This is connected to a form of realism, verism, which is the claim that the representation is necessarily retinal so we are shown what we see. What Wolf does so cleverly is to link this, via Dürer’s famous woodcut of a man drawing a female nude through a grid, to the idea of perspective as a form of possession. Although he does not pursue it, this connects gender, cartography, colonialism, and imperialism: to chart or draw is part and parcel of taking possession. Where then does this leave us? Description is not and cannot be neutral. Accepting that is not to abandon all hope of neutrality. I am not among the pessimists who have given up on the possibility of representation for versions of relativism. As I see the problem, it is not an all or nothing exercise. Recognizing that neutrality is impossible does not mean that it is not an ideal to which we should aspire (the double negative is intended). We are engaged in engagement. We are striving to comprehend, to explain and to display our comprehension, and to account for how we came by it.

This means that ekphrasis is an excellent model for anthropological representation in the widest sense, and particularly for thinking about how anthropology makes its object (to use Johannes Fabian’s [1983] subtitle), the relationships between the words we write (and the images we display), and the often distant, mythical, inaccessible other—the field, the source of our knowledge. Fabian argues that the use of the ethnographic present tense is used as an “othering device” in which the subjects of anthropological analysis are “othered” by being described in a timeless present. My argument is far more general than Fabian’s. At the root of my use of ekphrasis is the thought that, for most readers, the place which is my field is inaccessible. This is true whether or not the field is a remote island or down the road, “at home.” It may be easier for me to visit the site of Nigel Rapport’s fieldwork in Yorkshire in the north of England than that of Malinowski’s in the Trobriand Islands, but that gives no particular access to the specificities of his fieldwork let alone his collaborators, informants, and the social relationships on which his results rest. Ethnographic writing, even when as close to home as possible, reveals our distance from one another. In philosophy this is the “other minds” problem: how can we be sure what other people are thinking? (At the extreme this becomes solipsism: do other minds really exist?) This makes ethnographic accounts similar to those descriptions of lost statues which were the objects of the original ekphrastic statements. As readers of ethnography we have to try and work out what the author saw, what understandings their words are trying to convey, usually in the absence of any personal experience of the cultural milieu being discussed.

The idea of ekphrasis can be expanded in several different ways, for example, in the description of sound by analogy with the visual. The connection with anthropological analysis lies in the emphasis that ekphrasis puts on the aspect of translation in representation, moving from objects to words. At the core of ekphrasis is a representation involving translation across media from the visual or the aural to words. Further expansions are possible: for example, photographs index the moment of taking, but are taken to refer to the surroundings in ways which we know to be problematic. The contrast is useful; a photograph is a visual index to a visual scene, so it is not itself ekphrastic since no translation of medium occurs (but see discussion above as well as the next section for the relationships to the people depicted). There is common ground in the deictic anchoring of the indexation. An ekphrasis points beyond the words to the image it evokes, just as a photographic portrait points to the person in front of the camera. Without access to time machines we can never check, so the evidential validity—the warrant which justifies us taking a photograph as evidence— is less secure than it might seem at first. It is, I would suggest, frail in ways that resemble ekphrastic descriptions, where it is hard to judge their accuracy or success (in the absence of access to the image being described). So I suggest that thinking about ekphrasis is good for anthropology as a way of exploring the complexity of anthropological representation, not so as to abandon the task as impossible but to help us refine and nuance what we do (see Zeitlyn [2009a] on partial representations).

On that approach when giving a social science account we filter our analysis through theoretical concepts and construct (more or less) abstracted models of events in such a way that connections can be made to phenomena on the ground. Are these connections different in kind from the connections made between words on the page and the images they evoke? The concept of ekphrasis can accommodate the theoretical challenges of nonverbal data and the partiality of narratives. Ekphrastic social science can achieve visual and narrative sensitivity in an empirically responsible fashion.

Ekphrastic images in death

The painter and sculptor Michael Ayrton once painted a boy’s face peering through shards of broken glass—behind the unbroken sheet of glass of the picture frame (Contingency 1 or 2 c. 1955). The viewer wonders if the boy was playing and has broken a window, or if there is a deeper meaning: behind one translucent, frangible layer is another that cannot be broken. Can the boy leave the frame?

At the time of painting Ayrton was thinking about human frailty and arrogance; before producing his celebrated Minotaur series, he was reflecting on the myth of Icarus. My account of this painting is itself an ekphrasis and one which mirrors the classic form: after correspondence with Ayrton’s biographer I must report that the painting which I knew as a child is now of whereabouts unknown. The painting, even in my description, raises the question of whether any account can transcend itself (move beyond its frame). My answer is that it can but only with the collaboration or input of the readers/viewers to achieve deictic referentiality beyond the frame.

When Cameroonian citizens went to a local studio photographer to have photographs taken for inclusion on their National Identity Cards they were literally being inscribed into the nation-state. After a death, children may take the ID card of a parent to another photographer (or an anthropologist) to have the photograph (often the only image of a deceased elder) copied for display on a wall; doing so, they literally change the representation. Another case is the post-mortem drawing of a cross on the forehead of a photograph of someone who has died (see above). The images document a series of intentions about display, about the presentation of self.13 To write that much is easy, but it is worth considering that doing so is to represent other representations in a complex fashion and in a different medium—it is ekphrastic. I cannot make an assertion in a visual form, so I must annotate. As Barthes (1977) and Benjamin (1968) have argued (in different contexts), images are irredeemably textual.14

Clive Scott (1999) builds on that work in The spoken image, which examines the effect of captions and the ways in which narrative is achievable in still photography. Captions (and titles [ibid.: 49]) frame and hence change the image for the viewer. So I see the post-mortem annotations as challenging Scott’s use of Peirce’s distinction between index and icon (taking photographs as indexical and paintings iconic [ibid.: 40]). This returns us to ekphrasis. If the relationship between a photographic print and the person portrayed is indexical (in several senses of the word) then drawing a cross on the surface of the print marks a change in the relationship.15 It marks the print ceasing to be metaphorically ekphrastic and instead becoming ekphrastic in its classic sense: the only record of what the person looked like, it can no longer be compared with the original. In Peirce’s terms we can no longer trace back the chains of indexicality. The photograph is a trace which the viewer uses to construct (evoke) a person.16

Photographs change as their owners (possessors? possessees? subjects?) change; as they age and die. I have already discussed how in Cameroon it is common for photographs to be annotated after a death. As already mentioned, another common reuse is for an image to be taken from an album (or ID card), enlarged, framed, and displayed at postmortem funerary celebrations which may take place several years after the death (see Lancit 2010).

At the burial of my friend and colleague Rigobert Tueche (on 16th December 2000), everyone present knew how the photograph had changed; no longer a portrait of a proud young man, an academic at University of Yaoundé, preparing for his Doctorat d’Etat and about to go to France, his photograph was now a memento mori. The photograph is iconic. It changes, to misuse Peirce’s terms, from an index to a symbol since we can no longer link the image to its physical, three dimensional counterpart (or could no longer once the glass-fronted “open” coffin was been buried). As a map it ceased to work, just as his identity card, with its official photograph, now lacked validity.

Figure 2
Figure 2. Rigobert Tueche in the field 1998/9

The funeral was one of many acts—spread through the country and across two continents. The photograph was paraded around.17 The change was visible to all but invisible in the surface of the print. The corpse was viewed in the frame of the sealed coffin: a different sort of face behind the glass from that in the Ayrton painting—here, embalmed not painted, and at the moment of viewing profoundly different from Ayrton’s image, although now as I write, not different at all.

Long after the funeral in Cameroon I went to an antique shop in Rochester, Kent. In a box beneath a table were some photographs from the early twentieth century, perhaps eighty to one hundred years old. The likelihood is that those nameless people are already dead, as are many of the children also depicted there. But faces with names have a different weight from those in nameless images jumbled in a box for sale. Consider the question: Why have children? Some people say it is so their names will not be forgotten; that we live and reproduce to perpetuate memories of our existence. If this is right, then the archivist and the historian have something in common with the Church of Latterday Saints and their project to retrospectively baptize the entire world population into their church. To do this, Mormons believe, it is necessary to connect a living member of the Church of Latterday Saints to each individual, so they are baptizing a member of “their family.” Hence, their interest and support for genealogical research throughout the world. In effect—leaving the theology aside—they are putting names to faces. This is to link private and public spheres in ways that go against the difference between private and public uses that Berger (1980: 56) maintains—since the Church of Latterday Saints have presumably already baptized most of my readership by name into their Church without (most) of their permissions.

Figure 3. Rigobert Tueche funeral programme.

Knowing names does more than saying “look at this nice photograph.” It enables us to move from photograph as object (albeit a social object) to the life of people photographed. This is not always necessary; names are not always needed, so a photograph may evoke memories from people who recognize that they do not know anyone or any place depicted.18 However, if it shows people, places, or activities sufficiently close to those they did know, then it can serve as a prompt to memory and narration. Hence August Sander’s (anonymous) images now serve as documents for vanished ways of life, immensely evocative for all that we know none of their names (and similarly for many late-nineteenth-century or early-twentieth-century images). This caveat apart, knowing names takes us to specific people, participants in a particular complex of social transactions, and no longer “types” but individuals—no longer typical like the heroes of subaltern history, but known individuals placed once again (re-placed?) in the forefront of our attention as people. This is to return us to Jules Michelet’s “resurrectionalist history” which partly anticipated the work of Foucault. Michelet’s history, brings “the dead to life” especially by giving them names (more accurately: restoring knowledge of their names). That is why Berger (1980: 54) sees photographs needing memory (or for Barthes [1977], a caption) in order to achieve that “act of redemption.” Adding names exemplifies an ekphrastic dilemma. Strictly, saying for example, “this is a photograph of Huomnùàr” is ekphrastic only to those who know what Huomnùàr looks like, or least who he is, otherwise the ekphrasis does not work—it does not convey anything specific about the photographic image. We could say it suffers a failure of deictic anchoring, that indexicality has failed. Yet a caption, an annotation (often on the back of the print) giving a name opens the image to multiple future uses which its absence forecloses.

Anthropological representation can do the same: opening an account for future uses not imagined by the writer (this is to invert Fabian’s [1993] argument—less the timeless present robbing subjects of agency, more multiple, inchoate futures building on the past). To succeed in this is not straightforward. The challenge is how to be faithful to ones’ informants, to the conversations and experiences in the field and yet to enable reanalysis (at some possible point in the future); in short how to achieve an ekphrastic representation across time.

Readers are invited to reflect on how the discussion would work in the absence of the images discussed. This has recently been illustrated by Andrew Irving (2007: 198–202) in an essay about the photographs and accompanying commentary made by people with HIV/AIDS in Kampala. He has chosen to suppress some images for ethical reasons, but leaving the captions below blank spaces where the images would have been. As he (ibid.: 198) says, “the idea of ‘informed consent’ in anthropology is often tenuous and here I could not be certain that the people who were photographed were fully cognizant of some of the ways their images might be used or of their intended audiences. Thus instead of their pictures I show their absence and have provided the comments that Nalongo and Yudaya wrote to accompany the images they took.” Showing an absence might be a good way to describe a photograph or other memorial of a dead person.

The way in which a textual account can be read changes, and changes importantly, depending on whether or not it can be read alongside (possibly against) the images to which it refers. So ekphrastics, the multiple relationships of representation, form a fundamental part of our engagement with social science as we consider and discuss the relationships between analysis and data, explanans and explanandum (whether or not we use the term ekphrasis).


I have been arguing that anthropological representation, whether verbal or visual, is undertaken for specific purposes, and that it is ekphrastic—it points, evocatively, beyond the text or image. Moreover, it can be decomposed and the basis for its conclusions can be explained; it has empirical warrants from the exercise of fieldwork.

The idea of ekphrasis stresses the complexity of the relationship between representation and its object (the represented). When that object is lost or imaginary all we have is the representation. In anthropology we can play a game of “now we have access to both poles of the representational process, now we don't.” Never having been to the Trobriand Islands I can only work with Malinowski’s ethnographic reports (and those of the other ethnographers of the area, such as Annette Weiner). They evoke, ekphrastically, an image of Trobriand life. This is not to preclude the results of close and critical reading which can enable a variety of different readings to be sustained. In another case, relatively close to where Malinowski worked, Haidy Geismar (2009) has discussed how Layard’s ethnography—including his photographs—are understood in contemporary Malakula, almost a century after the original fieldwork. However, for this author, reading such work is considerably different from reading work from where my own ethnographic research is based. When I read what I and others have written about Mambila society, I can appreciate the simplifications and reductions which have taken place as part of the writing process, but I am still distanced from the people and relationships which grounded the fieldwork (as I am from other anthropology “at home,” see above). When I read ethnographies from colleagues working in the same area, I am in the middle ground—not as disadvantaged as when reading Malinowski, nor as privileged as when using material from the people I know best. I know enough to empathize with the writers, to grasp their points on the basis of knowing the sorts of things that go on; yet this knowledge may also provoke questions and an alternative telling. In all cases there is an inevitable gap between representation and the represented, and the relationship between them is ekphrastic. Understanding this is part of the process of bridging the gap and comprehending ethnography.

The suggestion that social science is ekphrastic helps us to recognize the importance of the relationship between representation and the represented, and the different types of engagement which characterize anthropology. Anthropological representations have many and various warrants which justify the partial and necessarily incomplete facsimiles we construct. Recognizing this, we can attempt to make more and better facsimiles, as well as to improve the ones we already have. Ultimately, ekphrasis points to the imaginative leap through the text or image to an understanding of what is portrayed. We must recognize that anthropological representations are inescapably ekphrastic.


Versions of this paper have been presented at workshops and conferences at the Pitt Rivers Museum, Oxford, at Wolfson College Oxford and elsewhere. I have greatly benefited from comments from and robust discussion with Marcus Banks, Elizabeth Edwards, Chris Morton, and Richard Vokes as well as some early discussions with David Reason from the University of Kent which were as inspirational and challenging as ever. The reviewers and editorial comments from Hau have also helped me to clarify the argument, and Anna Rayne helped sharpen the prose and challenged me to be clearer about my intended meanings. The remaining faults are all my own.


Asad, Talal, ed. 1975. Anthropology and the colonial encounter. London: Ithaca Press.

Bajorek, Jennifer. 2013. “Decolonizing the archive: The view from West Africa.” Aperture 210: 66–69.

Barthes, Roland. 1977. Roland Barthes. Translated by Richard Howard. London: Macmillan.

———. 1981. Camera lucida: Reflections on photography. Translated by Ricahrd Howard. New York: Hill and Wang.

Benjamin, Walter. (1936) 1968. “The work of art in the age of mechanical reproduction.” In Illuminations: Essays and reflections. New York: Schocken Books.

Berger, John. 1980. About looking. London: Writers and Readers.

Brown, Penelope B. and Stephen C. Levinson. 1978. “Universals in language use: Politeness phenomena.” In Questions and politeness, edited by Esther N. Goody, 56–290. Cambridge: Cambridge University Press.

Bruner, Jerome. 1995. “The autobiographical process.” Current Sociology 43: 161–77.

Buckley, Liam. 2006. “Studio photography and the aesthetics of citizenship in the Gambia.” In Sensible objects: Colonialism, museums and material culture, edited by Elizabeth Edwards, Chris Gosden, and Ruth B. Phillips. Wenner-Gren International Symposium Series. New York: Berg Press.

Carrier, David. 1987. “Ekphrasis and interpretation: Two models of art history writing.” British Journal of Aesthetics 27: 20–31.

Edwards, Elizabeth. 1999. “Photographs as objects of memory.” In Material Memories, edited by Marius Kwint, Christopher Breward, and Jeremy Aynsley, 221–36. Oxford: Berg.

Edwards, Elizabeth and Janice Hart. 2004. “Introduction: Photographs as objects.” In Photographs objects histories: On the materiality of images (material cultures), edited by Elizabeth Edwards and Janice Hart, 1–15. London: Routledge.

Fabian, Johannes. 1983. Time and the other. New York: Columbia University Press.

Geismar, Haidy. 2009. “Stone men of Malekula on Malakula: An ethnography of an ethnography.” Ethnos 74: 199–228.

Goffman, Erving. 1959. The presentation of self in everyday life. New York: Doubleday Anchor Books.

Grice, Herbert Paul. 1975. “Logic and conversation.” In Syntax and semantics III: Speech acts, edited by Peter Cole and Jerry L. Morgan, 41–58. New York, NY: Academic Press.

Hacking, Ian. 1983. Representing and intervening: Introductory topics in the philosophy of natural science. Cambridge: Cambridge University Press.

Hanks, William F. 1990. Referential practice: Language and lived space among the Maya. Chicago and London: University of Chicago Press.

Hauser, Kitty. 2007. “Tracing the trace.” In Shadow sites: Photography, archaeology, and the British landscape 1927–1955, 57–104. Oxford Historical Monographs. Oxford: Oxford University Press.

Heffernan, James. 1991. “Ekphrasis and representation.” New Literary History 22: 297–316.

Herzfeld, Michael. 1997. Cultural intimacy: Social poetics in the nation-state. London: Routledge.

Irving, Andrew. 2007. Ethnography, art, and death. The Journal of the Royal Anthropological Institute, 13 (1): 185–208.

Jurkevich, Gayana. 1999. In pursuit of the natural sign: Azorín and the poetics of ekphrasis. Lewisburg: Bucknell University Press.

Kripke, Saul. (1972) 1980. Naming and necessity. Cambridge, MA: Harvard University Press.

Lancit, Matthew (director and producer). 2010. Funeral Season (ou La Saison des funérailles): Marking death in Cameroon. DVD.

Latour, Bruno. 2002. “What is iconoclash? Or is there a world beyond the image wars?” In Iconoclash, edited by Bruno Latour and Peter Weibel, 16–40. London: MIT Press.

Martin, Edwin. 1986. “On seeing Walton’s great-grandfather.” Critical Inquiry 12 (4): 796–800.

Marcus, George and James Clifford, eds. 1986. Writing culture: The poetics and politics of ethnography. Berkeley: University of California Press.

Mitchell, William J. T. 1994. “Ekphrasis and the other.” In Picture Theory. Chicago: University of Chicago Press.

Mustafa, Hudita N. 2002. “Portraits of modernity: Fashioning selves in Dakarois popular photography.” In Images and empires: Visuality in colonial and postcolonial Africa, edited by Paul S. Landau and Deborah D. Kaspin. Berkeley: University of California Press.

Parmentier, Richard J. 2009. “Troubles with trichotomies: Reflections on the utility of Peirce’s sign trichotomies for social analysis.” Semiotica 177: 139–55.

Quine, Willard O. 1960. Word and object. Cambridge, Massachussetts: M.I.T. Press.

Ramos, Manuel J. 2004. “Drawing the lines: The limitations of intercultural ekphrasis.” In Working images: Visual research and representation in ethnography, edited by Sarah Pink, Laszlo Kurti & Ana I. Alfonso. London: Routledge.

Rose, Rex. 1999. “The last days of Ernest J. Bellocq.” Exquisite Corpse: A Journal of Letters and Life, (10). Accessed August 30, 2014. http://www.corpse.org/archives/issue_10/gallery/bellocq/index.htm

Ruby, Jay. 1991. “Death and photography: A review essay.” Visual sociology 6: 82–86.

Ruby, Jay. 1995. Secure the shadow: Death and photography in America. Cambridge, MA: M.I.T. Press.

Rugg, Linda H. 1997. Picturing ourselves: Photography and autobiography. Chicago: University of Chicago Press.

Sacks, Harvey. 1995. Lectures on conversation. Volumes 1 and 2 (in one volume). Oxford: Blackwell.

Said, Edward. 1978. Orientalism. New York: Pantheon Books.

Scott, Clive. 1999. The spoken image: Photography and language. London: Reaktion Books.

Silverstein, Michael. 1985. “The functional stratification of language and ontogenesis.” In Culture, communication, and cognition: Vygotskian perspectives, edited by James W. Wertsch, 205–35. Cambridge: Cambridge University Press.

———. 2003. “Indexical order and the dialectics of sociolinguistic life.” Language and Communication 23: 193–229.

Stevens, Wallace. 1997. “Description without place.” In Collected poetry and prose. Library of America. New York: Knopf and Penguin.

Taussig, Michael T. 1980. The devil and commodity fetishism in South America. Chapel Hill: University of North Carolina Press.

———. 1993. Mimesis and alterity: A particular history of the senses. New York: Routledge.

Vokes, Richard, ed. 2012. Photography in Africa: Ethnographic perspectives. Woodbridge: James Currey Publishers.

Walton, Kendall. L. 1984. “Transparent pictures: On the nature of photographic realism.” Nous 18 (1): 67–72.

———. 1986. “Critical response: Looking again through photographs: A response to Edwin Martin.” Critical Inquiry 12 (4): 801–8.

West, Shearer. 2004. Portraiture. Oxford: Clarendon Press.

Williams, Malcolm. 2003. “The problem of representation: Realism and operationalism in survey research.” Sociological Research Online 8 (1). http://www.socresonline.org.uk/8/1/williams.html

Wittgenstein, Ludwig. 1953. Philosophical investigations. Oxford: Blackwell.

Wolf, Bryan. 1990. “Confessions of a closet ekphrastic: Literature, painting and other unnatural relations.” Yale Journal of Criticism 3: 181–203.

Worth, Sol. 1981. “Pictures can't say ain't.” In Studying visual communication, edited by Larry Gross. Philadelphia: University of Pennsylvania Press.

Zeitlyn, David. 2004. “The gift of the gab: Anthropology and conversation analysis.” Anthropos 99: 452–68.

———. 2005a. “Introduction.” In Joseph Chila and Samuel Finlak: Two portrait photographers in Cameroon, edited by Ingrid Swenson. London: Peer.

———. 2005b. Words and processes in Mambila kinship: The theoretical importance of the complexity of everyday life. Lanham, Maryland: Lexington Books (Rowman and Littlefield).

———. 2006. “Visual anthropology and properties of the medium (or The visual anthropologist in the digital library: From filmstrips to salient stills and back to Barthes).” Visual Anthropology Review 21: 3–13.

———. 2009a. “Understanding anthropological understanding: For a merological anthropology.” Anthropological Theory 9: 209–31.

———. 2009b. “Archiving a Cameroonian photographic studio with the help of the British Library ‘Endangered Archives Programme’.” African Research and Documentation 165: 13–26.

———. 2009c. “A dying art? Archiving photographs in Cameroon.” Anthropology Today 25: 23–6.

———. 2010a. “Representation/self-representation: A tale of two portraits.” Visual Anthropology 23: 398–426.

———. 2010b. “Photographic props/The photographer as prop: The many faces of Jacques Tousselle.” History and Anthropology 21: 453–77.

———. Forthcoming. “Redeeming some Cameroonian photographs: Reflections on photographs and representations.” In The African photographic archive: Research and curatorial strategies, edited by Darren Newbury and Christopher. Morton. London: Bloomsbury Publishing.

Zeitlyn, David and Roger Just. 2014. Anthropology in between (realism and relativism). Newcastle: Cambridge Scholars Publishing.

Les Antinomies de la représentation: Anthropologie d'un procédé ekphrastique

Résumé : Cet article s'intéresse à une question anthropologique majeure: quel est le lien entre la représentation et le représenté? Qu'est-ce qui motive ou justifie leur inévitable déconnexion? Ce serait une erreur de ne pas prendre au sérieux la représentation sous prétexte qu'elle est erronée, oppressive, ou parce qu'elle prête à confusion. La représentation fait partie de la cognition en général et du langage naturel en particulier. Pour cette raison, elle est inévitable et elle fait partie de la manière dont nous pensons et parlons du monde. En invoquant tour à tour des éléments d'anthropologie visuelle et linguistique, je propose l'hypothèse que les photographies et les portraits constituent une riche archive à partir de laquelle on peut penser les différents types de mandats impartis aux représentations anthropologiques. La conclusion générale de ce propos est que la représentation anthropologique peut être conçue comme une forme d'ekphrasis (un récit ou l’évocation d'une image ou d'un objet typiquement non-présent) fournissant le lien indexical et déictique entre la représentation et l'objet représenté. De même que la “similarité engendre la différence”, “la représentation engendre l'ekphrasis.”

David ZEITLYN has been conducting field research in Cameroon since 1985. He has worked on sociolinguistics, traditional religion, endangered languages, and more recently on Cameroonian studio photography. He gave the 2003 Evans-Pritchard Lectures at All Souls College Oxford on the life of a senior Mambila woman, Diko Madeleine.

David Zeitlyn
Institute of Social and Cultural Anthropology
School of Anthropology and Museum Ethnography
51 Banbury Road
Oxford, OX2 6PF
United Kingdom


1. Further discussed in Zeitlyn 2005a, 2009b, 2009c, and 2010.

2. See Mitchell (1994), Jurkevich (1999) and Hefferrnan (1991).

3. Patterns of usage on Twitter may give the lie to Worth’s claims: users tweet images to refute other tweets which may be textual statements or text and image.

4. I note that in a different context (life writing) Jerome Bruner collapses this distinction (1995: 167).

5. In mid-2014 only the two men at Mɔgɔ’s feet were still alive.

6. A referee asked why I call this “annotation” and not “defacement.” In all the cases I have seen the marks are carefully positioned so faces can still be seen, even when the mark is on a forehead. The intention seems to be to add “information” not to reduce it. Contrast with the celebrated case of the literally defaced photographs of Ernest Bellocq (see Rose 1999).

7. Elsewhere Roger Just and I (2014) explore what Herzfeld (1997: 165 ff) calls the “militant middle ground” in a longer treatment.

8. Michael Silverstein (1985: 217–25) presents a Peircean approach to this topic. Silverstein’s Table 1 (218–19) lists the distinctions needed to analyze semiotic functions. Some might feel that distinguishing “dicent indexical Legisigns” from “rhematic Legisigns” (True Reference from Attributive Reference), is too much like butterfly collecting. I find such neologizing classifications off-putting and although the distinctions of the second third, between icon, index, and symbol have been widely adopted, it is telling that the tenfold path of the three thirds (Parmentier 2009) has not been pursued outside a small circle of Peircean semioticians. Indeed, I note that Richard Parmentier (ibid.: 147), one of Peirce’s champions, says that “the complexity of his trichotomies … does not offer a proportional degree of direct assistance for the work of social analysis.”

9. Anthropologists are hence in a sort of “double-bind” or a relation of bad faith, caught between their engagement with their informants (see Zeitlyn and Just 2014) and their theoretically sophisticated colleagues in philosophy. It is not our job to criticize the naïveté of relatively unschooled informants on the basis they haven't read Western philosophy, yet as academics we must demonstrate to our peers that we have read Quine, Putnam, and Wittgenstein, and can play the citation game to accrue the symbolic capital that comes from dropping the right names.

10. Description is intimately related to realism. Realists have to deal with indexicality to link their descriptions with the world in ways that nonrealists do not. Non-realists have no truck with concepts like “the world” so there is nothing for them to describe.

11. As a referee to this paper points out some forms of metapragmatics are common: “Don't talk that way” or “She told the story angrily” are metadescriptions that we use frequently. While accepting this, I would also note that to complain about the style is often to avoid dealing with the substance so a choice has been made.

12. See Scott (1999: 25) for a parallel discussion of Barthes’ refusal to show us the photograph.

13. Goffman (1959); see Buckley (2006) and Mustafa (2002) for African examples from Gambia and Senegal as well as essays in Vokes (2012) and in a special issue of History and Anthropology, e.g. Zeitlyn (2010b).

14. As is argued elsewhere (Zeitlyn 2006) this is also relevant to the indexing and management of image collections.

15. Here it is appropriate to acknowledge Elizabeth Edwards’ path-breaking work on both the physicality of photographic prints and the way they have been used in anthropological arguments and as objects of anthropological study. See, e.g., Edwards (1999) and Edwards and Hart (2004).

16. This is close to archaeology in Collingwood’s (not Foucault’s) sense. Another powerful metaphor is Rugg’s (1997: 238) use of scarring—in which a print as scar evokes the event (wounding) which produced it. See also Hauser (2007: 102–3).

17. Cf. Ruby (1991 and 1995), who discusses American varieties of mortuary and funeral photography, and West (2004: 65), who discusses antecedents in the larger historical tradition of portraiture on death and commemorative representation.

18. I am grateful to Elizabeth Edwards for alerting me to this possibility in her comments to an earlier version of this paper presented at the Photographic history workshop, Pitt Rivers Museum, Oxford, December 7, 2011.