Politics of Photography


Mihai Surdu

Collecting photographs, assembling numbers, suspecting Roma

“The camera is integrated into a larger ensemble: a bureaucratic-clerical-statistical system of »intelligence«. This system can be described as a sophisticated form of the archive. The central artifact of this system is not the camera but the filing cabinet.”

Sekula, 1986:16

There is general agreement among scholars and Roma activists alike that the label Roma covers a wide range of very diverse groups. These groups hardly come together under a common denominator: “This imagined community shares no common language (only a small minority speaks one of the dozens of often mutually unintelligible dialects of Romani), culture, religion, identity, history or even ethnicity” (Kovats, 2003).

Yet the current approach of Roma inclusion by a politics of difference, to borrow from Epstein (2007), appeals to photographic representations of a Roma collective. Displayed by international organizations in policy papers and websites, photographic representations intended to portray a Roma collective oversimplify, stigmatize and sometimes racialize while ethnicising poverty (Surdu, 2016). I further explore the depiction through photography and categorization in censuses as two distinct and yet related practices of grouping Roma.

Making a Roma population visible is not a new project; for a long time, a “Gypsy” population was in the focus of experts. Therefore, I aim to reflect here on the goal of making Roma visible, on its techniques, assumptions and rationales as well as on its relationships with expert practices of producing populations, such as censuses. My argument runs as follows: Since there is no coherent Roma group and individual identities are in constant flux, a Roma group cannot be accurately represented photographically as a whole. Nevertheless, in the past, photographic collections aiming to depict “Gypsies” (and later Roma) as a group were assembled. Roma were also counted by censuses which were qualified in their modern beginnings as “true” representations of the “social body”. I argue that, despite the irreducible diversity of the umbrella group, a common racialized vision of Roma people, embedded in both photographs and data, relates photography to censuses as tools for producing “Gypsy/Roma groupness”.

I hope to demonstrate using the examples below that “Roma photography” and “Roma numbers”, both of which draw on police methods of identification, are joined by an expert vision which informs and is informed by a collective vision. The punctum (Barthes, 1980) of the vision shared by photography and census is race, understood as a “relational object” created through scientific practices and situated “beyond fact and fiction” (M’Charek, 2013).

Photography has been a companion to physical anthropology since the beginnings of the entangled histories of the two fields in the mid-19th century, as it was deemed an ideal tool to capture race in distinct, measurable, recognizable and unchanging racial types. As such, photography was employed as a tool of racial classifications based on anthropometric measurements of body parts and skin color as well as contextual clues of clothing, objects denoting occupations, housing and other aspects of life.

To give one example, the Swedish State Institute for Race Biology (SIRB) photographic collection assembled a “Gypsies and their descendants” album as part of a large (12000 photographs) racial taxonomical long-term project sorting three main races (“Nordic”, “East Baltic” and “Lappish”) and “Jews”, “Walloons”, “Gypsies” and “mixed types” as minor races (Kjellman, 2014). Pictorial methods owing to both anthropometry and social documentary traditions were joined in the SIRB eugenicist photographic project, which continued until the late 1950s according to Kjellman (2014). Photographs were believed to make an unmediated reality legible, since photography was regarded as a nonintervention technique par excellence, being beyond human manipulation (Barthes, 1980; Sontag, 1973).

However, at the turn of the century, scientific cultures started to diversify the visual strategies used to lend credibility to clustering humans in racial groups. Physical anthropologists and human geneticists built their visual persuasion strategies of portraying human diversity by adding more abstract visualizations such as “maps, tables, drawings, diagrams, family and phylogenetic trees” sometimes in addition to photographs (Lipphardt, 2015). Yet photography has apparently not ceased to be considered a tool of mechanical objectivity, despite trained vision being a pre-requisite of sampling, staging, recording, assembling and interpreting ethnic/racial photographic profiles.

The other project of ethnic clustering performed by the census, which appeals to numbers for making groups visible, also required a set of activities before arriving at “objective” ethnic numbers: persuasion of their subjects, census manuals, lists of ethnicities and codes, instructions, coding and post-coding categories1.

Era Trammer | Politics of Photography | Non Fiction | Austria, Germany | 2018 | pho_00054 Rights held by: André Raatzsch — Erika Trammer | Licensed by: André Raatzsch — Erika Trammer | Licensed under: CC-BY-NC-ND 4.0 International | Provided by: RomArchive The images cited in the film were kindly provided by: György Buzás Jánonsé Buzás József Buzás Sándor Buzás Palkó Lászlóné Mária Raatzsch Jürgen Raatzsch György Stalter Jánosné Turkacs Rozália Treiber Chad Evans Wyatt Burt Glinn – Agentur Focus / Magnum Photos Josef Koudelka – Agentur Focus / Magnum Photos Hungarian Film Office - Magyar Film Iroda Museum of Ethnography – Budapest Fortepan – Online Photo Archive Florian Schuh – Picture Alliance Rhein Neckar Zeitung

Despite the impossibility of depicting a Roma group (or any other collective for that matter) in photographs, visual representations of a “Gypsy/Roma collective” are historically present and real. This can be seen in photographic collections of “Roma” or “Gypsies” carried out for different reasons and using various types of expertise: criminological, eugenicist, physical anthropological, ethnographic, medical, journalistic and by the development industry, amongst others.

Since photographs intended to depict Roma as a group are spread over so many fields of expertise, with different logics and practices, I do not intend to substantially reflect on such photography across those fields in this paper. With examples taken from physical anthropology, forensic photography and censuses, I am problematizing racialized visions of groupness in relation to Roma.

I attempt to answer the following questions: What visions were and are enacted in assembling a Roma collective through photography? What are the epistemic assumptions and conditions of possibility in which Roma-related photographic collections arise? What role does trained vision have in making “Roma photography” appear objective? What effects does “Roma photography” have?

Suspecting Roma through Photography and Censuses

Police methods of identification, as recounted by Sacks (1972), rely on being able to infer possible crimes from “suspicious persons”. “Populations” of “suspect appearance” are created by police evaluating the “moral character” of persons based on their visible traits. Based on such “suspect populations”, the task of the police could accordingly be described as being “able to use their appearances to isolate candidates for investigation” (Sacks, 1972). Even today “suspect populations” are the object of police investigation: police stops and checks of identity of minority members in Europe, including Roma, are much more frequent than for other people (FRA, 2010).

In part, the ethnic profiling of Roma by the police is rooted in historical expert practices of objectification and quantification through photography and censuses. To be sure, turning Roma into a suspect population requires also a collective vision to be effective - not only police expertise. As Sekula (1986:9) notices, photography when circulated by the police was seen as instrumental in curbing criminal “careers” by dragging “wider citizenry in the vigilant work of detection”.

In the early history of photography, photographs were taken or used by the police in their work of identification. Starting in second half of the nineteenth century, photography was used by the state for documenting, surveilling and controlling an increasing number of mobile people (Sontag, 1973). Perhaps the first use of photographs for making itinerant people a “suspect population” was in Switzerland, as early as 1852-1853. This earliest available collection of police photography, containing 220 portraits of itinerants, is hosted by the Swiss Federal Archive.

Photography in this context was used to document the identity of people lacking Swiss citizenship - the group surveilled by the police would later be recognized by the name Jenische (Meyer and Wolfensberger, 1998) and considered by some scholars and policy experts as part of an umbrella Roma group. Some of the photographs in this earliest experiment in judicial photography aiming to settle vagrants and investigate them as a group (Jäger, 2001) used captions of occupations (e.g. basket weaver, watch maker) and objects of particular occupations (a basket held by a photographed woman) to link individuals to groups by representing occupational branches. It is important to highlight here the similarities between police visions of a “Gypsy/Roma collective” and some academic and administrative practices of sorting and counting people (Surdu 2016).

Interview Mihai Surdu | Non Fiction | 2018 | pho_00281 Rights held by: André Raatzsch — Era Trammer | Licensed by: André Raatzsch — Era Trammer | Licensed under: CC-BY-NC-ND 4.0 International | Provided by: RomArchive The images cited in the film were kindly provided by: Simone D. McCourte / World Bank Photo Collection

As my next example reveals, assembling a collective does not necessarily need photography to make groupness around visible appearances; it is enough to have a trained vision when doing censuses. Moving to the Kingdom of Hungary in 1893, the first special census on “Gypsies” carried out by statisticians, demographers and ethnographers, assisted by the police, relied on a racialized vision of sameness and difference informed by the racial science and in turn contributing to it. According to this vision, the census had to be carried out using observational methods, meaning that persons resembling the “anthropological character” of a “Gypsy” counted as “Gypsies’”.

As the census takers had no calipers or other measurement tools with them, presumably the “racial trait” of a darker complexion coupled with “nomadism” were good enough indicators of “Gypsiness”. Though the 1893 census’s rationale was a long-standing concern with “Gypsy” nomadism, those categorized as vagrants by the census were a bit more than three percent more than those recorded (Johnson, 1998; Havas, 2002). However, various people at that time had professions requiring travel and were not labelled nomads or “Gypsies”; therefore, visual appearance was a key prompt for inclusion in the “Gypsy” group.

Police identification practices and physical anthropology were intertwined projects informed by 19th century racial science seeking to classify people based on physical characteristics. As Sekula (1986) argues, police photography historically draws its techniques and assumptions from physiognomy and phrenology, the same sources which contributed to physical anthropological photography. Strangers to the police were the Other of the anthropologist: they were scrutinized by the same expert gaze on the body and, more importantly, on the face. Among other body measures, skin color was widely used as the indicator of a racial type, including by anthropologists studying and photographing “Gypsies”. These anthropologists were focused more on capturing “nomadism” (e.g. horses and tents), clothes and other contextual details (as for example by Swiss anthropologist Eugene Pittard, also a president of the Gypsy Lore Society who photographed “Gypsy nomads” in Romania).

In parallel to police-led censuses of “Gypsies” (e.g. the 1895 French national census of “nomads, bohemians, and vagrants” and 1905 “Zigeuner” census in Bavaria which included photographs from police records), physical anthropological collections of “Gypsies” were part of a long-term scientific project. On longue durée, police-led censuses of “Gypsies” relied on an expert consensus (supported in the past as today by scholars) that “Gypsies” are undercounted by regular censuses and that neither self-declaration nor the language spoken are reliable indicators for recording a “Gypsy/Roma group” (see Surdu and Kovats, 2015; Surdu, 2016). Accordingly, state authorities and scholars considered (and some still consider) police methods of identification (e.g. visual inspection, grouping by occupation, socio-economic status, ghetto residency etc.) more reliable, thus contributing to making “Gypsies” and later Roma a suspect population hiding behind false identities.

The logic of suspicion put into practice made it possible to assemble what may be the largest collection of police-anthropological photography intended to pin down the identity of “mobile populations” for surveillance purposes. Together with anthropometric measurements converted into an abbreviated notation system translating body measurements into text, the so-called portrait parlé breveted by Bertillon for identifying “habitual” criminals, photography became a widespread forensic technique documenting criminal records by the end of the 19th century. Bertillon’s anthropometric system coupled with a standardized police portrait (face and profile) was intended for individualization and deconstruction of the criminal type for practical purposes of classification and policing (Sekula, 1986). Yet, despite intended individualization, bertillonage helped to prolong police practices of profiling for “criminal types” (Hagins, 2013). Two decades after the introduction of bertillonage, the new police method of identification was applied for othering, identification and surveillance of people and families practicing itinerant occupations.

As part of controlling a variable “dangerous class”, family ID cards named carnet anthropométrique d'identité, nomades were introduced in France in 1912 as proof of a collective identity (special identity documents were requested in France for gens de voyage until recently). Police photography was an indicator of the dangerousness of traveling people, especially those lacking citizenship but also of political enemies, anarchists and revolutionaries, impoverished peasants or members of the proletariat changing jobs and places. Part of those identified with nomad IDs according to 1912 French law were sent to labor camps a few decades later and others were deported to Nazi Germany. Similar special ID documents were issued in 1927 in Czehoslovakia for “Gypsies”, many of whom would become victims of Nazi extermination policies based on racial classification.

Acknowledging here that police vision and practices of categorization were essential to the social formation of a “Gypsy” group2 does not mean to deny the broader range of expertise and political attention which currently keep Roma in focus, nor to remain distant to the history of the present.3

My last example refers to a recent census, the 2011 Romanian census which had a record number of non-answers to the optional question of ethnicity. I interpret this refusal to identify with an ethnic label as an act of resistance to the state's categorization practices, that is a refusal of “seeing like a state” (Scott, 1998). In contrast to my interpretation, some social scientists in Romania (and part of the public commenting online) turned the findings of the census into an occasion for ethnic labeling: in their opinion, the more than 1200000 people who did not answer ethnicity questions were Roma trying to hide their »real« identity, thereby treating them as an ethnically suspect group. This expert interpretation supported by a political-expert consensus that Roma are massively undercounted by censuses is based on a racialized vision of ethnicity (Surdu and Kovats, 2015): visible characteristics, social status and living in a segregated neighborhood are its main candidates, much the same for police and anthropological photography as well as police-led censuses of “Gypsies”.

Concluding remarks

I recall a recent conversation about Roma and photography with a friend. He told me: “As a child, when I was browsing family albums with my parents, they used to say ‘This is one of your grandfathers.’ They then went on to talk about his life events and personality. I don’t remember even once them saying ‘This is your Roma grandfather.’”

In family albums, photos of relatives do not come with an ethnic label. My friend and his grandfather are in a relationship unmediated by any form of groupness except that of being part of the same family.

Yet contemporary Roma identity politics further deepen societal divisions, solidify racialized visions of Roma groupness and ironically risk fueling racism through discourse and policies intended to be anti-racist (Kovats, 2003; Tunali, 2011).

Acknowledging that creating visibility for marginalized ethnic or racial groups is not dividing camps among progressives and conservatives, Phelan admits strength in unmarking difference:

“There is real power in remaining unmarked; and there are serious limitations to visual representation as a political goal. Visibility is a trap [...]; it summons surveillance and the law; it provokes voyeurism, fetishism, the colonialist/imperial appetite for possession. Yet it retains a certain political appeal.” (Phelan, 2005, 6)

I have tried to argue that photographic collections depicting Roma and assembling the “true” numbers of Roma have, since their inception, been objects of an expert gaze being used by entangled disciplinary projects in population making. Photographs featuring Roma, often collated from photographic clichés and complemented by numbers, stabilize identities that are skin-deep and hence fictitious, although real in their social formation and consequences.