'Lean on Me': Sifarish, Mediation & the Digitisation of State Bureaucracies in India

Through an ethnographic focus on Muslim neighbourhoods in a North Indian city, this article traces the effects of increasing digitisation of Public Distribution Systems (PDS) and ID provision in India by examining the implications for relations between the state, low-level political actors and local populaces. The article explores the practice of sifarish (leaning on someone to get something done) which, it is argued, cannot be seen within simplistic rubrics of ‘corruption’ but instead comprises a socially embedded ethical continuum. With one of the stated aims of digitisation being the displacing of informal mediation, the ethnographic material illuminates the efforts of low-level political actors to navigate emerging digital infrastructures. Digitisation, however, does not end mediation and carries with it ideological, political and economic interests. This, the article argues, enables state/people spaces of mediation to be commodified and marketized and further cements processes of marginalisation experienced by India’s Muslim minority.


Introduction
On the 1 st of July 2015 the Indian Prime Minister, Nerandra Modi made a speech at an event titled 'Digitising India' in which he stated, 'digitalisation of all government work is essential to deal with problems like corruption, help provide transparent and efficient governance and bridge the rich-poor divide' (The Telegraph India 2015). The speech formed part of an ongoing process which, as with increased use of ICT in various parts of the world, aims to 'escape the limits of the old paper state' (Breckenridge 2014: 16). Digitisation projects in India have come to embody notions of good governance, transparency, modernity and anti-corruption. They are also imagined as a means to address marginalisation and social exclusion through providing direct connections between people and the state (Rao & Greenleaf 2013).
Yet, ethnographic and other research has illuminated the unpredictable ways in which digitisation plays out. In poor neighbourhoods of Delhi, for example, fingerprint readers utilised within India's biometric registration system could not detect the prints of homeless people due to lifetimes of ware on the flesh (Rao & Greenleaf 2013;cf. Rao 2018). During fieldwork for this article, in the provincial North-Indian city of Saharanpur and surrounding areas, similar issues were reported amongst manual labourers and agricultural workers.
Additionally, the use of iris scanners had triggered rumours of blinding and mind reading in some outlying villages. The intersection of human and non-human -within India and elsewhere -often means that digitisation is not a quick fix solution to address marginalisation and/or lack of inclusion but an (often messy) assemblage of bodies, technologies and algorithms which situate digitisation within broader 'sociotechnical systems' (Lowrie 2018).
Simultaneously, the digitization of state bureaucracies are not detached from broader economic, political, ideological and social processes but fold into pre-existing forms of privilege and exclusion (Graham 2005) and remain deeply embedded in social relations (Rao & Nair 2019).
By focusing on the Muslim mohallas (neighbourhoods) and villages surrounding Saharanpur, this article traces the effects of increasing digitisation of Public Distribution Systems (PDS) and ID provision (such as the newly introduced biometric Aadhaar card). These bureaucratic and material transformations are bound up not only in the construction of the Indian state but also within global processes of technological and ideological change (cf: Lowrie 2018). The article focuses on implications for everyday relations between the state, low-level Muslim netas (politicians) and pradhans (village heads) -serving as representatives in Saharanpur's Nagar Palika Parishad (City Council) or in nearby Gram Panchayats (Village Councils) -and those they represented. Here, the emphasis falls onto the practice of sifarish (leaning on someone to get something done), the exercise of which was key to legitimising the authority of local netas and other political actors.
Against this background, the article makes three primary arguments. Firstly, that sifarish cannot be conceptualised within a simplistic rubric of 'corruption' but sits within an ambivalent ethical continuum. Sifarish acts as a conduit in processes of claim-making between marginalised people and the state (albeit in ways that are cross-cut by embedded social relations). Whilst similar arguments are reflected in earlier ethnographic work from India and elsewhere (Gupta 1995;Singerman 1995;Ruud 2000;Smith 2001;Mitchell 2002 Chatterjee 2017; Jayal 2019; author 2020a), digitisation not only undermines more informal conduits of connection between marginalised people and the state but also interplays with the ideological positioning of a nationalist politics that is increasingly focused on the establishment of a 'Hindu Rastra' (Hindu Nation) (Harriss et al, 2017) and with a post-liberalisation economic landscape that foregrounds neoliberal models of development (Münster & Strümpell 2014).
Secondly, the paper shows that, contrary to many official representations, the material transformations heralded by digitisation are no less ambivalent than are pre-existing, socially embedded networks. Whilst various continuities persist, the material transformations that comprise digitisation of ID and PDS provisioning are altering spaces of mediation between people and the state. Rather than dis-embedding mediation from the social and the political, however, digitisation carries with it the ideological and political environments in which it is deployed and becomes entangled in interplays at local, national and international scales. At a more theoretical level, and in a context where some strands of assemblage theory -through the dilution of human agency and de-reification of 'the social' -(cf. Latour 2005) have moved towards a distinctly organic representation of relations between humans, technologies and other non-human actants, this article demonstrates the continued relevance of political economy within bio-technological assemblages (cf. Fine 2005).
Finally, the article illuminates how digitization can become a carrier of hegemonic market interests by enabling the state/people space of mediation to be commodified and marketizeda form of accumulation by dispossession. Here, I argue that the peopled infrastructures through which mediation pathways operate (Chaudhuri 2019;Elyachar 2010;Simone 2004;Singerman 1996) constitute a distinctly spatial context. As Doreen Massey (1992) has argued, space is constructed through social interrelations but "…the social is [also] inexorably also spatial" (p.80). Given these spatial properties, and despite highly blurred boundaries (Gupta 1995), spaces of mediation offer -as with urban space, spaces of culture, spaces of nature, spaces of artisanal or peasant production and non-capitalist space -a frontier for forms of (primitive) accumulation. It also leads to degrees of displacement (albeit in ways that are incomplete, contested and uneven) of the peopled infrastructures within which sifarish and related practices are embedded, These transformations, I contend, act as a conduit which further intensifies the draining of power from the Muslim minority population. As David Harvey (2007: 23) argues: The creation of […a] neoliberal system has entailed much destruction, not only of prior institutional frameworks and powers (such as the supposed prior state sovereignty over political-economic affairs) but also of divisions of labor, social relations, welfare provisions, technological mixes, ways of life, attachments to the land, habits of the heart, ways of thought, and the like.
The article is based on ethnographic fieldwork conducted in 2016 and 2017. This involved spending time with various actors and 'hanging out' at sites of mediation such as ration stores, government bureaus and the offices of agents and brokers. Additionally, 50 interviews were conducted with the public and a further 40 with mediators including netas, pradhans, social workers, ration dealers and agents. The article also draws on from my decade-long engagement with the area (cf. author 2018; 2019; 2020a) and from material developed through research-led teaching in the UK (author 2020b). The text is arranged into five sections. The first, provides an ethnographic vignette to orientate the reader. The second situates ongoing digitisation programs against a background of intensifying marginalisation of India's Muslim minority. This is followed by a section tracing shifts from a 'paper' to a 'digital' state at both structural and everyday levels. I then turn to providing a detailed account of sifarish, its everyday enactments and the interplays between the long-standing peopled infrastructures in which it is embedded and digital transformations. Finally, I offer ethnographic and broader material to trace how 'spaces of mediation' provide sites for accumulation through dispossession and marketisation.

Digital Dilemmas in the Mohalla
I had known Usman 1 for eight years. His house, larger than most in the neighbourhood, was situated in a crowded mohalla in the north of Saharanpur. Usman, a sturdy man in his mid-40s of a kind and welcoming disposition, lived with his wife, two daughters and three sons. The eldest son was married, with one child, and his daughters were studying at college level. His second son, in his early 20s, worked with his brother in the family's furniture making business, consisting a small workshop and a showroom. The final son was somewhat younger and attended a local English Medium School. As a local ward-level 2 neta, Usman held sway within the community he represented and often cultivated social networks and support in the narrow gullies (lanes) that constituted the area.
Like many netas, his position enabled him a degree of access to the state and influence within local government and civil infrastructure. For Usman, and those he represented, utilising sifarish to lean on officials in order to ease the passage of documentation and assist in obtaining forms of identification or PDS provisioning was envisioned as an obligation to voters and community rather than as 'corruption'. Thus, Usman's ability to perform sifarish was central to legitimising his position as a local neta. Voters supported him on the basis that he would mediate their interactions with the state. Whilst many netas had a reputation of failing to follow up, being able to do so -or at least engage in performances which displayed 'potential' to do so -were key illustrations of power and authority in the eyes of others.
Recently, however, online processing of applications for cards, documents and paperwork had reconfigured mediation pathways and impinged upon performances of sifarish, as Usman describes: People always come to my house. If someone comes and says they need a ration card, then I should help as I am the neta. I can do this because in the [government] department I have respect. People can use me to put pressure on officials, or they can use money [he laughs]. When the officers see money they make no mistakes [in the application]. In every department everyone takes a bribe. If they do not have money, then they must ask me to lean on that official.
Often people complain that I am not helping them. That person [referring to a local man waiting to speak with him] is saying "yaar tumne mera kam nahi karaya" [phaa! You did not do my work] … but how can I make a card for him when the [web] site is closed. Before the computer system we could do these things with our sources but now it's difficult […] Educated people understand the computer system but here many are uneducated and cannot understand why we are not able to make the system work. They blame me and say that I am not doing their work, they think that sifarish is needed but cannot see that the computer does not listen to any pressure. If the system for applications is closed so it is closed! Now everything depends on the computer system and not on sources. Only when the site is open can the feeding [of the computer] start.
Waseem was a neta from a neighbourhood close to Usman's. A little older and considered more senior within local circles, he experienced similar problems when the digital Ration Card Management System (RCMS) was introduced. Unlike paper processes, utilising sifarish to hasten an application had become challenging as officials themselves were rendered impotent: If the website is open, then we do not need to worry about it. We just call the officers and at once the work is done. However, now the website is closed, so we are waiting as even the officer can do nothing. When it is open there is no tension as I know all the officers very well. I have been a neta here for 40 years, so everybody knows me. When everything was on paper then usually I could do the work but now it is not always possible.
Usman's and Waseem's narratives illustrate the ambiguities and tensions experienced by low- and digital forms of bureaucratic governance, challenged their ability to perform sifarish, and thus risked the undermining of their authority in the eyes of others. On the surface, then, these online systems had done exactly what was intended by dis-embedding applications from the local social context, rationalising the system and giving ordinary people direct connections to the state. As the following sections of this article show, however, digitisation -whilst creating degrees of transformation -did not eliminate state-people mediation and remained deeply embroiled with the marginalising processes experienced by Indian Muslims.

Indian Muslims, Identity, Digitization & the State
In the context of Saharanpur's Muslim mohallas, as from many Muslims elsewhere in India 2), a vision of a plural Indianness that was articulated by its proponents in the independence movement as 'composite nationalism' (Madnī 2005(Madnī [1938). give non-Muslim 'illegal' migrants from Afghanistan, Pakistan and Bangladesh access to citizenship on the basis of assumed religious persecution whilst excluding Muslim 'illegal' migrants from the same countries (Bhat 2019). The implication being that Muslims whowhether through actual immigration to India or simply due to a lack of paperwork -cannot prove their right to citizenship will likely face differentiated consequences from non-Muslims.
The CAA and NRC represent very material -in the sense of being legitimised within the law -enactments of graduated citizenship along communal lines, but are also symptomatic of the Fundamental here, is the Aadhaar Card, the world's largest biometric ID system launched in 2009 (Rao & Greenleaf 2013). This involves creating a massive citizenry database housed within the Central Identities Data Repository (CIDR) (UIDAI 2019). Aadhaar -along with numerous other projects within India's state bureaucracy -is intended to increase inclusion, address identity gaps, monitor civil servants, rationalise welfare, and circumvent informal channels of brokerage and mediation (Bhatia & Bhabha 2017). Hence, the digitisation of India's PDS and ID provision is often imagined as an "…(anti-political) force that is able to penetrate the layers of mediation […,] layers inhabited by the various middlemen, intermediaries, brokers, fixers, and even politicians who […] encumber the relations between state and society" (Noy 2014: 108). This is not limited to the Aadhaar card alone. Thus, Ration . It is to this longer lineage of contributions -which span both paper-based and digital contexts -that the following section now turns.

Bureaucratic Materialities: From Paper to Digital States
Debates on bureaucracy have long involved tensions between seeing bureaucracy as symbolic of growing social complexity, rationality and development, and questions of discipline, authority, coercion and (class) power (Marx 2011(Marx [1867Arendt 1967;Gramsci 1971;Weber 1978;Foucault 1980). Ethnographic material, however, has illustrated that power lies not only in bureaucratic structures and the actions of bureaucrats but also in mundane everyday interactions. Here, bureaucracies may render the claim-maker or citizen passive by, for example, making them wait (Jeffery 2010; Auyero 2012; Hoag 2011), by performing degrees of (often racial, gender and class-based) indifference (Hertzfeld 1992), via the fetishization of systems, procedures and documents (Gordillo 2006;Folch 2012) or through structural violence (Gupta 2012). Yet, we also understand the messiness, negotiations and contestations at intersections between states and populations (e.g. Corbridge et al 2005;Mathur 2016; author et al 2019; Truelove 2020); a juncture which is often blurred, ambiguous and occupied by multiple actors (both human and non-human) who dissipate notions of the state as a clearly identifiable, homogenous entity (Mitchell 1999;Gupta 2005;Mathur 2017).
Documents can reinforce state power and render the state present in the homes, workplaces and pockets of citizens (Scott 1998). Paper documents and online processes can act against individual agency, rendering citizens and others as passive 'patients of the state' by, for example, producing temporalities of marginalisation (Auyero 2012). Simultaneously, the ability of certain forms of documentation to embody the state can make them desirable. Gaston Gordillo (2006), for example, describes how a campaign to be recognised as citizens enacted by the indigenous Argentine Chaco community led, when successful, to the fetishization of state documents as symbolic of belonging. Gordillo, however, reminds us that fetishization was produced in a historical context of violence and exclusion, thus blurring the subject/subjected dichotomy and emphasising the interrelations of material artefacts with historical and political projects.
The mediating qualities of documents, cards and paperwork can, however, be inverted in various ways. Fake documents, for example, may embody the aura of the state to the same degree as their authentic counterparts (Srivastava 2012). Paperwork -such as Right to Information Requests -when deployed by activists may reverse disciplining process, turning the lens towards government bureaucrats (Webb 2013;Mathur 2016). Citizenship, and its associated documentation, may be refused -as with many Tibetan refugees in India -in order to maintain claims to sovereignty elsewhere (Mcgranahan 2016). Whilst variegated, clearly documents and paperwork are key mediators between populations and the state. Matthew Hull (2012a) illustrates this in the context of urban planning in Islamabad (Pakistan), which he sees as being constituted and negotiated through 'a regime of paper documents' (p.1). Here, 'order and disorder […] are produced through the ceaseless circulation of millions of maps, forms, letters, and reports among bureaucrats, politicians, property owners, Imams (prayer leaders), businessmen and builders' (p.4). Hull points out that, as with other forms of human and nonhuman mediation, the role played by documents is often invisibilised, a process which is itself a 'tactic of power' (2012b: 253; cf. Mathur 2017). As with Michael Foucault's (1977) deliberations on 'power' more generally, for Hull, the power enshrined in state documents and cards is often ambivalent and can be both oppressive and productive (at times simultaneously).
On occasion such enactments of power might be explicit -the attempts to make legal distinctions between Muslim and non-Muslim 'illegal' migrants in the CAA and NRC, for example -but they may also become entangled with apparently mundane acts. A common complaint in Saharanpur concerned the mis-spelling of Muslim names by Hindu officials. This could have long-term implications lasting years or even lifetimes. In a nearby town I met Rizwan, a school teacher in his mid-40s with three teenage sons. Rizwan had invested in their education and wanted to open opportunities for them to study abroad. When applying for his eldest son's passport, however, past errors in various unrelated documents created a series of problems: I went to get my son's passport with his birth certificate and a letter from the school.
They asked also for my wife's ID like a ration card but I did not have it so I had to get our marriage certificate. This was in Urdu so I got it translated into English and stamped by the court. No one had noticed, though, that her name on my son's birth certificate was spelt wrong (Asna instead of Asma), so I had to get it changed which needed a bribe. Finally, I got the new certificate but they said they had to refer to the health register where my wife's name was again misspelt. I then had to go and get a certificate from the hospital where my son was born and only then did I get the passport.
We also had many problems with the ration card. Some clerk had not written the names of my family members but instead just written famous Muslim names like Akbar and Aziz. There must have been a survey years ago, but the surveyors couldn't be bother to visit our house and just wrote some Muslim names.
Government officials can never learn our names. My mother is so ashamed as her name is Amir Bano but in all her official documents she has the male name Amir Khan (also a famous Bollywood actor). In the Aadhaar card and ID card the spelling may be different. When the officers see this they may reject our cards. It is the officer's mistake. Hindu officials always have problems writing our names.
If we check it straight away we can stop them otherwise nothing can be done later.
Issues surrounding accidental or intentional mis-spelling of Muslim names had not disappeared with the arrival of digitisation but instead became bound up in interconnections between online systems and paper documents, with many errors transferred to online databases. Thus, through the intentional and unintentional acts of bureaucrats and others, marginalising forces experienced by Muslims in the city continued within processes of bureaucratic transformation and were not diminished through digital 'rationalisation'. As Ian Lowrie (2018)  clientelism and other practices, that move beyond simplistic frameworks of 'corruption'. As Arild Ruud (2000) suggests, such a reductionist approach '…prevents us from understanding that these are practices developed within a fully mature normative system of no less moral validity than any other normative system' (p.271) (cf. Smith 2001). Writing on Pakistan, Nasir Islam (2004)  For netas in Saharanpur, displaying these capacities was key during election periods, with the capability to obtain PDS and ID provision for others often central during canvassing. However, complaints about broken promises made by netas keen to win votes were commonplace, as Mumtaz, an elderly woman in one of the mohallas exclaimed: The neta of our area filled this form for my pension. He does this so that he will get votes but I have not heard anything for a long time, and [now that he is elected] he will not answer my requests.
Forms of exclusion often become configured around local and familial politics. Farida, a recently divorced woman, described how her former husband's relationship with a local neta led to him blocking her from obtaining a ration card independent of her ex-husband's family. interfere and there are rules and quotas. People say, "you made promises to us, it was fake promises, so I will not give you my vote in the future". But in every department the system is difficult. Government programs are full of tension for us. This is the biggest problem for us, always there is political pressure.
This was a common narrative, often discussed in terms of dual pressure from constituents and the state. In a village 8km outside Saharanpur, Sajid, the local pradhan described '…for some people you cannot say no [to performing sifarish] as you have links to them. Also, they put pressure through voting, so then it is hard to refuse'. The divergence between promises and what Pradhans, netas and others could deliver meant that they had to develop strategies to legitimise their authority even in the context of failure. Amongst other actions, this involved the public berating of officials. Although a change in outcome may not result, the neta had at least displayed their authority in the eyes of their constituents (see: author et al 2019).
Yet to reduce sifarish to instrumentality would negate more complex motivations. Usman (whose story opened this article) saw sifarish not only as a means to get things done but as a moral duty. Sitting in his home, we discussed the recent Food Security Act (FSA). Introduced in 2013, the FSA was to provide subsidised grains to Below Poverty Line (BPL) households, but it's rollout proved controversial with accusations of vote buying (Banerjee et al 2014) and of negative impacts on farmers (Peschard 2014). For Usman, however, his concern surrounded area-based quotas which did not account for variability on the BPL population numbers 4 , an uneven distribution that required an ethical fix (cf. Rao 2018; Pia 2017). Without the ability to influence policy, the primary tool available was sifarish, which he described using to help people bypass the quota. After explaining this, Usman leant across the table and, in a sincere tone, said 'this is not a fair system as poor areas loose out. If I get a family registered [with sifarish] then this is my duty'. Sifarish, then, is normatively constructed and understood within its own localised moral and ethical frameworks. Along with other sets of relations and practices, it also forms a deeply embedded web of infrastructural conduits through which people/state interactions take place. The final empirical section of this article turns to the processes embedded within digitisation that enable degrees of commodification of these informal spatialised infrastructures and details the implications this has for marginalised Muslims in Saharanpur's mohallas.

The Commodification of Mediation Spaces
At the opening of this article I argued that sites of mediation should be considered in spatial terms. This is significant for understanding the transformations taking place in Saharanpur and elsewhere as peopled infrastructures mediated (for better or worse) through socially embedded political networks become, to degrees, marketized and through accumulation by dispossession.

In his work on HIV and AIDS prevention and control programs (funded by Word Bank loans)
in Pakistan, Ayaz Qureshi (2015) has illustrated how these programs became commodified through neoliberal models of development which foregrounded market-driven, over statebased, forms of delivery. This, Qureshi argues, created a context where World Bank money was primarily paid to international NGOs and private sector businesses or consultancies, rather than making the 'risk groups' -who were the stated target of the programs -the main beneficiaries. This, Qureshi suggests, leads to the "dispossession of many for accumulation by some in the ambiguous space created by the reorganization of AIDS bureaucracy" (p.46) In discussing the UK government's digitally-based Transparency Agenda (TA) -a system through which the public, the press and others can monitor and 'witness' the expenses claims and salaries of MPs, senior civil servants and other officials -Penny Harvey et al (2012) show how TA not only led to new ambivalences but also created the material conditions -an 'entrepreneurial' space -within which private sector mediators could flourish. In this case, the activities of these new mediators consisted of producing Apps and software to manage or catalogue data and to allow the general public and others to do the work of transparency monitoring. Whilst Harvey et al's case suggests a degree of 'newness' to the space that emerges, Qureshi's descriptions elucidate a reconfiguration of pre-existing bureaucratic infrastructures and other sets of relations. As David Harvey (2007) articulates, new spaces of accumulation under neoliberalism are often not 'created' but instead involve the displacement of that which occupied the space previously. This 'creative destruction' is, for Harvey, a primary mode through which the commodification of space occurs. Qureshi (2015), Harvey et al (2012) and other rejoinders (e.g. Pattenden, 2010;Waquant, 2012Schwiter et all, 2018 have illustrated how bureaucratic transformations within welfare systems and development programs can be carriers of market interests. Simultaneously, these, and others works, also show that while the state may, at times, be undermined, it remains deeply embroiled in these processes. Additionally, neoliberal forms of self-making (which embody entrepreneurialism, competitiveness, independence, self-reliance etc.) are shown to penetrate beyond structural processes and become entwinned in subjectivitities of low-level bureaucrats and political actors. In Saharanpur, too, an 'entrepreneurial' space (albeit one characterised by informality, precarity and insecurity) had emerged within a mediation infrastructure that was increasing colonised by the private sector. Aadhaar applications and other digitised services, for example, were now primarily carried out by small-scale, local, 'private sector' agents who, in a context of low literacy and IT skills, mediated (for a charge) between citizens and the state's online apparatus.
Saharanpur, as with many other Indian cities, has seen a boom in these outlets in recent years with a large semi-formal industry developing around the administering of online processes.
Many of these operations run within existing businesses such as internet shops or grocery stores, but various new outfits have also sprung up leading to the opening of a myriad of small agencies. These brokers, who according to government figures number around 900,000 nationally (UIDAI 2020a), need to be licenced by the state and are required to undertake training (either at a local training centre or a 'Mega Training Camp', often facilitated by private providers) and must sit a short exam to obtain Unique Identification Authority of India (UIDAI) certification (UIDAI 2020b). Together with the supply and manufacture of equipment by companies such as Morpho and Mahindra Satyam, along with a myriad of other public and private sector actors undertaking a variety of roles, according to the UIDAI's own language, this assemblage creates an 'Aadhaar ecosystem'. The bureaucratic hierarchy therein comprises (1) Registrars (either government or private sector) who can register citizens but also orchestrate registration to subcontracted (2) Enrolment Agencies. In turn, agencies can establish permanent or temporary (3) Enrolment Centres comprising one or more (4) Enrolment stations where a photograph, iris scan, ten fingerprints and personal details are registered on equipment bought or leased from hardware and software vendors for upload to the Central Identities Data Repository (UIDAI 2019;pp. 1947-1948.
Aslam, who ran an enrolment station as a private sector agent in a corner of his father's shop, sat close the bottom of this structure. The long-standing business that he and his father operated offered Photostat services, train tickets and online passport applications. A year before, he had invested in the equipment for Aadhaar-issuing, at a cost of around 1 lakh 5 , and paid to sit the exam. Technically, Aadhaar Enrolment and mandatory updating is free of charge at Aadhaar Enrolment and Update Centres, but most private sector outlets charge an additional fee of around 100-150rs paid by the applicant. Whilst some agents, such as Aslam, purchased the equipment privately, others were provided it when they signed up to one of several private sector companies who would take commission on each transaction.
In a nearby mohalla I met Naseer, who had previously worked for Smart ID. Aged only 21 and with limited education, he occupied a very different positionality to the netas and others who generally dominated spaces of mediation. He had been recruited through a broker, who in turn received a commission for each agent they contracted, a model which has been proclaimed by some as embodying the ideals of individual entrepreneurialism (e.g. Prahalad 2004). The involvement of various private sector companies led to the emergence of a complex, and increasingly crowded, card and document issuing market in which companies competed, not only with each other, but also with local netas keen to retain their position within newly emerging digitised infrastructures. Naseer described how, in 2011, there had been only one company, Smart ID. However, their operations were suspended after a State Bank of India (SBI) sponsored Aadhaar Camp, set up on a field outside the city to issue cards, was hijacked by local netas intent on claiming the initiative for themselves in order to legitimise their authority in the eyes of others. 'During the camp', he described, 'there was a lot of competition between the different netas, they all tried to show their power, but they did it without the permission of the SBI, so the program was stopped' 6 .
When the program of subcontracted Aadhaar agents restarted in 2013, Smart ID were joined by a variety of competitors in what was now a rapidly expanding market. Inevitably this created increased competition and impacted on the sub-contracted agents themselves, as Naseer described: In the beginning, a company could get 50rs for each card they made. Now though it is only 15-20rs. In one city there may be 8 to 10 companies working. Now the boys who do this work are divided into groups and then themselves make different distributors [e.g. a pyramid scheme arrangement]. The work is divided according to area, but some groups cannot complete enough so they go back to other work. Now the system for camps is closed, so if you want to make the Aadhaar card you have to pay. In the camps it was free but in the permanent centre the fees can be anything from 50-1000rs.
With the camps no longer running, many of those seeking an Aadhaar card found themselves attending private centres such as Aslam's or going through agents such as Naseer. Here, the fees could vary substantially. To get a card legitimately (or at least under circumstances where you have all the ID and information needed) the price was rarely more than 150rs. However, for those who lacked the necessary details and documents this could rise substantially to cover the 'risks' of completing online documentation without all the required checks. Naseer explained how he began to engage in the taking of bribes whilst undertaking agent work: At first, we refused to help those people who did not have all the ID and relevant things. However, our groups were mostly young boys and we wanted some enjoyment like movies or entertainment. Many people came for making the Aadhaar cards illegally, but we selected those people who could give us good money, who looked rich, we only did this with this sort of person. In Saharanpur people gave 1000rs for it but in Dehradun 7 we could get up to 5000rs. We could complete this work in just 5 minutes.
The digitisation of Aadhaar and other PDS schemes, then, has not removed forms of 'corruption' or mediation from the process but has often involved the marketization of mediation pathways and actors, with workers such as Naseer operating in precarious conditions best described as 'organised informality'. Here, the symbolic terrain of formality (e.g. uniforms, training programs, digital technology etc.) act to mask the inherent informality, insecurity and precarity of labour (Gooptu 2013). Hence, through the processes outlined in this article, spaces of mediation within a context previously dominated by socially embedded political actors, were opened up to the private sector and to emerging forms of accumulation within PDS and ID provisioning systems.
Whilst poorer and marginalised communities in India have long experienced the state through the market in the form of both non-monetised (e.g. votes or familial/community obligations) and monetised (e.g. bribes) transactions (Tarlo 2003), the marketization of PDS and ID delivery has reconfigured mediation pathways in favour of corporate interests. This is far from the disembedding implicit in many representations of digitisation. Mediation becomes monetised at all levels and involves new actors (both human and non-human). Yet, mediation itself has not disappeared and there is no direct connection between people and the state.

Conclusion: Digitisation at the Margin
It is important to clarify at the opening of this conclusion that my intention has not been to reify sifarish as an idealised mode of mediation. As we have seen, it is a highly ambivalent practice which creates various forms of exclusion and inclusion. However, what can be argued is that emerging digital pathways are no less ambivalent. Whilst the netas and others discussed in this article play to their own concerns and seek to legitimise their authority, they are also embroiled in degrees of obligation to the community from which they emerge. Thus, digitisation has resulted in Muslim residents of the mohallas losing a degree of access to the state through their 'own' intermediaries. Newly emerging actors, such as Aslam and Naseerdescribed in the latter section of this article -may also be drawn from the city's Muslim mohallas but they lack the political clout and ability to deploy sifarish that local netas possess.
Thus, they have limited capacity to modify and correct perceived inequalities or errors in bureaucratic systems. They are also, often, motivated by short-term monetary gain rather than the long term accumulation of social capital (in the Bordieuan sense of the term) that represents the primary drive for netas and others mediating between their vote base and other sites of political and bureaucratic engagement.
Consequently, for India's marginal and minority communities, this results in a draining away of power from political actors emerging from within the community to new players who operate in a precarious neoliberal market economy which negates, to a degree at least, an identity-based politics of representation. However, the 'spaces of mediation' discussed in this article are by no means fixed and, as illustrated by the hijacking of the Aadhaar camp by various netas, bureaucratic transformations are in constant process, embody forms of continuity as well as change and are regularly contested and negotiated in various ways. As the digitised state becomes more embedded, so these low-level political actors may find new means through which to legitimate their authority or prove capable of capturing emerging digital pathways within existing avenues of patronage, clientelism and sifarish. As the materiality of the 'paper state' gives way to the supposed 'immateriality' and 'rationality' of online systems, so our attention as ethnographers should turn to both intended and untended consequences of material transformations at the level of everyday politics and sociality. Yet, clearly these are not coincidental, organic formations assembling together in contexts where human agency and 'the social' are reduced to equivalents of non-human actants. Rather, what this analysis of digitisation shows us is the centrality of political economy, ideological projects and social relations within the assemblage of digital transformations. 1 All names are pseudonyms. 2 Saharanpur comprises 70 wards. Wards may constitute more than one mohalla. 3 Writing and research for this article was undertaken prior to Covid-19, with the pandemic arriving in the latter stages of revision. It is worth noting, however, that the processes described here have also become deeply entangled within the biopolitics of the crisis. India's Muslim minority, in particular, have experienced an intensifying of state violence and both right-wing media and political actors have attempted to connect the spread of the pandemic with India's Muslim population under the banner of 'Covid Jihad'. 4 There has recently been a lot of attention given to this in the media with celebrities and politicians being exposed as having access to subsidised rice despite their substantial wealth. For example, the recent scandal involving Deepika Padukone, Sonakshi Sinha, Jacqueline Fernandez and Rani Mukherjee (The Hindu, 2016). 5 100,000 INR 6 Naseer's account is his own recollection so may contain some inaccuracies. However, the general context he describes is aligned with broader fieldwork and research conducted for this article. 7 Uttarakhand's state capital around 60km from Saharanpur.