Knowledge, numbers and the Northern Territory Intervention: re-conceptualising facts in remote indigenous Australia.

Clark, Christian

FACTS, LIFE AND THE NORTHERN TERRITORY INTERVENTION

Let us begin in the remote North East Arnhem Land community I call Ngunhili, (2) as the Intervention is getting underway.

   Between the shop, clinic, and council buildings is a shaded grassy
   area. It is a place to sit, congregate, wait for lifts to
   homelands, assemble work crews, and from which to participate in
   the main routines of settlement life: shopping, banking, government
   business. It is the early dry season, and a beautiful time to be
   outside.

   On the grass, Centrelink, the agency responsible for social
   security payments, has set up a table at which to conduct
   interviews registering locals for Income Management. The Centrelink
   staff each has his or her own laptop, and wireless Internet that
   connects them to the Centrelink database. A Yolnu (3) woman sits with
   one of them. On the table is an A4 sheet of paper. Printed on it
   are clipart pictures of food, a building and some other icons, each
   with an 'equals' sign to the right and then a line inviting
   inscription.

   It is being explained to the woman that for each of her children
   fifty dollars per fortnight will go to the school for their meals.
   This totals one hundred dollars for this mother of two and '100' is
   written above the line. 'So you have one-hundred-and-eighty-four
   dollars left over for the shop or saving' says Barb, the Centrelink
   employer. 'Maybe you can send one hundred dollars to the shop and
   save eighty-four?' The mother replies softly, but my ears, attuned
   to numbers, hear 'one-hundred-fifty'. 'That means you are only
   saving thirty-four dollars!' exclaims Barb. There is a pause. A
   silence. Barb looks determined. The mother looks uncomfortable and
   harassed. 'There is a phone call for me in ten minutes,' she says,
   and gets up and walks away.

   Centrelink is also setting up a BBQ. This is the carrot for drawing
   people to a meeting about Income Management. That Centrelink has
   closed the shop for an hour is the stick, but it does not appear to
   be working. A bigger stick might be needed. 'Just tell them we'll
   cut their payments if they don't turn up,' jokes one Centrelink
   staffer. The Centrelink boss confidently reassures her colleagues
   and a few others standing around like me: 'People just don't
   understand, but when their payments are cut, that usually brings
   them out of the woodwork.' A local Yolnu man standing next to me
   says (less loudly) that it is not that people don't understand--in
   fact they do understand very well--but Income Management is being
   imposed on them. 'So what is the point,' he reasons, 'of hearing
   about it again and again?' At noon, the work crews break for lunch
   and save the Centrelink staff from having to eat all the sausages
   themselves. 'Well, that was a waste, they all have jobs,' complains
   the Government Business Manager. She continues: 'What I don't
   understand is, if there are houses with so much money going into
   them, how come people are starving?' The question is a complaint: a
   conclusion more than an introduction.

Why did the woman walk away? How does the State remain certain even in confessing its confusion? A common account of this episode would focus on politics, with the Yolnu woman offended and the State certain it will prevail. Another common telling would focus on knowledge, with the Yolnu woman confused about the numbers and calculations and the State sure in its rational administration. Can there be another account which is not so quick in its attribution of knowledge, and hence power? Can our analyses help create spaces in which Yolnu can be considered more knowledgeable in their actions and the State less hegemonic? This paper develops such an account by re-conceptualising knowledge and politics together.

The question/complaint ending this episode exemplifies the mystification expressed by the State, its administrators (and others) when confronted with remote Indigenous livelihoods. These ways of living are precisely 'what we don't understand'. Moreover, it is the experience of this mystification as immobilising that is particular to the liberal State which is committed to rational and harmonious progress. (4) In State-speak, remote Australia has been described as a 'failed State', (5) reminding us of the moral framings embedded in understandings and their attribution of blame. The divisions that scaffold these framings--between law and custom, fact and myth, who is modern and who is savage--are hardened in remote Australia. The most recent and dramatic State reaction to this mystification has been the Northern Territory Emergency Response, or the Intervention.

During the 2007 Federal election campaign, the incumbent conservative Liberal-National Coalition bewildered its opposition and the Australian public with announcement of an immediate federal government take over of 73 Indigenous communities in the Northern Territory (NT). The Coalition referred to a recently published report concerning child abuse in the NT entitled 'Ampe Akelyernemane Meke Mekarle--Little Children are Sacred' (6) as evidence of an unexpected and critical emergency, although the report's authors explicitly recount the government's inaction toward many such reports published in the preceding decades. (7) The Coalition ignored all recommendations of the report, and instead pursued an immediate and sweeping overhaul of legislation with respect to Indigenous Australians, their rights to land, employment, schooling and social security. This was 'unmatched by any other policy declaration in Aboriginal Affairs in the last forty years' (8) and a radical break from the previous policy of self-determination. With this mobilisation of only a few ministers and senior bureaucrats taking charge of hundreds of Centrelink staff, the army and an unlimited budget, Indigenous Affairs Minister Mal Brough and Prime Minister John Howard hoped to evade the tedious work of democratic governance (and campaigning for re-election) by proclaiming a moral crisis. Despite the election of a different government, and a one-year review recommending drastic alterations to the Intervention, (9) it has continued for the most part unchanged.

Three episodes of the Intervention inspire and structure this paper. These stories do not present the facts of the Intervention. As a non-Indigenous researcher, I am not writing for those who are affected by the Intervention, nor do I write about the Intervention from some position outside the situation of the Intervention. My interest is in how researchers such as myself can engage with the situations such as the Intervention in our research. The episodes I present here are ethnographic, experimental puzzles around facts and the Intervention. They are ethnographic in that they are informed by fieldwork in a remote Indigenous community in Arnhem Land, and because they attend to the many practices of the Intervention, rather than the many perspectives of it. (10) They are experimental in that they generate new accounts of these empirical goings-on, including generating new tools of analysis. (11) And, finally, they are puzzles. Rather than consigning our problems to the mystification of an 'other' and seeking solutions through 'our' modern knowledge, I accept that stories tell of both problems together with their solutions. (12) I will focus therefore on the extraordinary yet mundane routines of how one measure of the Intervention, Income Management, was implemented. This may seem odd, even trivial, against the grander narrative of the politics of State, colonialism and race. However, that story has been told, and while the one presented here is less familiar, it is no less important.

INCOME MANAGEMENT AS A MORAL/RATIONAL IMPASSE

While the Intervention appeared as extraordinary, in one significant way it was not. Elisabeth Povinelli argues that the separation of a 'moral sensibility' of what is right and a 'critical rational knowledge' of what is right is characteristic of the modern liberal state. (13) In this case, good governance (consultation, thorough policy development and clear implementation) and good deeds (caring for children) appear separated. Povinelli argues, however, that this 'difference is interesting only and exactly because of the way it works as a generative impasse in the liberal discourse about the institutions of cultural recognition.' (14) In this paper, I argue that the Intervention measure of Income Management worked as a form of generative impasse in the institution of the market economy.

By telling episodes of the implementation of Income Management in a remote community, I tell of Income Management disrupting what we usually accept as fact--numerical representation, in this case of money-value. This disruption provokes an analysis in which facts, and in particular numbers, never seem to be achieved as the objective, value-free entities that they aspire to be. My response to seeing the way the Intervention troubles numbers as performing as usual (that is, as singular, value-free objects) is to understand numbers not as objects but as routinised practices. This account of facts captures the generative nature of Povinelli's impasses, which she sees as making possible new forms of collective life. This account of number also avoids the mystification (and hence exclusion) of Indigenous agency, knowledge and livelihoods.

The Intervention, and particularly Income Management, disrupted much more than the fact of numerical representation. It severely affected the people living in the 73 communities prescribed by Minister Brough, and any friends and relations visiting these communities from the many outstations they service, or regional centres. These people wore the intimidation, uncertainty, mistakes and incompetence of the Intervention's implementation. Income Management quarantined half of all payments from Centrelink to people living in prescribed areas. The intention was to prevent this money being spent on alcohol, cigarettes, pornography and gambling. Despite initial efforts, (15) employed people--including those employed by the Community Development Employment Programme--were not able to have their incomes quarantined.

The implementation first allowed shops to hold quarantined payments in accounts if proprietors did not allow this money to be used on proscribed items. One community store used paper tokens and a hole punch, another recorded account balances on index cards, and another had a voucher system. Later Centrelink facilitated the use of store cards from major retailers, none of which operated in the communities affected. Approximately one year after Income Management began, an ATM-like card, called the Basic Card, that could be used to purchase food, clothing and medication at registered stores was issued to all people on Income Management. The administration of Income Management cost $88 million in its first year, $105 million in its second, and legally required the suspension of the NT Anti-Discrimination Act and the Commonwealth Racial Discrimination Act (1975), an action maintained despite strong criticism by a United Nations Rapporteur in 2009. (16) It also included a 'sub-measure' for which Centrelink staff could 'take action' and quarantine fifty dollars per fortnight per child from a parent's social security payments for a School Nutrition Program. The Program receives payments all year round irrespective of school holidays and other times of justified absence, providing lunches only for attending students (and often staff) on school days. (17)

In North East Arnhem Land, the Arnhem Land Progress Association (ALPA) had already developed a Foodcard. A community Women's Centre initiated the Foodcard in 2004 as a voluntary card to assist people in managing their money, specifically for nutritious food. It was funded by the Commonwealth Government, and the accountancy firm Deloitte Australia was to implement it. (18) ALPA's nutrition policy specifies unhealthy items, which would not be purchasable with the Foodcard. The Intervention was announced in the months before the Foodcard became available, and this card, originally conceived as voluntary, was designated as the recipient of people's compulsory quarantined social security payments. Moreover, the local board of the ALPA store decided that for Ngunhili, the proscribed items (tobacco, cigarettes etc.) would be extended to include ALPA's list of unhealthy items.

CLIPART, KIDS AND CASH: NUMBER AS REPRESENTATION

In the opening episode, a woman receiving a social security payment sits at a table with a Centrelink employee. She is shown a document that uses clipart pictures of all the different institutions her money is to be sent to: the school, the shop, the bank (savings), the home (cash). The fact that she is the parent of school-age children prompts the substitution of her two children with two amounts of fifty dollars, which together make one hundred, and so the Centrelink employee utters 'one hundred' and writes down '100' to the right of the equals sign and the picture for school. The remainder of the payment is calculated, one-hundred and eighty-four, and again this needs to be distributed between the remaining institutions: the bank and the shop. 'One-hundred and fifty' is uttered by the woman as an amount for the shop. This is not immediately accepted and the interview collapses, leaving the unfinished form on the table and the woman 'unmanaged'.

This little routine of registering a person for Income Management is simultaneously one of assembling and distributing social security payments. One cannot be an income managed customer without having an income that is managed. The routine is primarily carried out through numbers, which are crucial in differentiating a total payment into separate amounts that can be distributed to an already differentiated set of institutions. Numbers here work as representations of the amount of newly differentiated amounts of money. Puzzling about representation has a long history in philosophy and the philosophy of science. I will use the work of Ian Hacking to tease out this practice of working numbers as representation.

Hacking defines representations not as statements, words or images (as is often asserted in structuralist or semiotic accounts of sign and referent), but as objects, visual or not, crafted and 'intended to be more or less public likenesses'. (19) ('Pictures in the mind' are not representation as they are not publicly or collectively produced and/ or witnessed.) (20) In other words, '100' is not a representation of an ideal number, or a particular pile of money. For Hacking, likeness emerges first as a character of the object-as-representation itself.

The category of the 'real', in this account, comes only after the crafting of representations. In the presence of multiple representations, different metaphysics--what constitutes the 'real' and how we know it--are mobilised in order to 'sort good systems of representation from bad ones.' (21) Hence, metaphysics becomes most often contested in the presence of already-made representations. Numerical representation is very seldom seriously threatened by any other alternative representation and hence there is little consideration of the metaphysics it affects. Number, as a form of object (that is 1, 2 and 3 are all number, as a dog, cat and owl are all animal), seems to remain ephemeral, floating somewhere in the domain of practice that is most often overlooked. Hence, in most situations numbers remain as real entities, however momentary, with a self-evident unqualified likeness. How might we analyse objects such as numbers, and their potential goodness, as they are being made? To do this we must interrogate the work of crafting numerical representations in practice.

Through the routine of registration, Centrelink employee and customer, pen and paper, gestures and utterances, the written '100' and uttered 'one hundred' attempt to be produced as public likenesses. Centrelink is very familiar with this public likeness number and accepts it as constitutive of the routine of registration. In doing so, the representation '100'-drawn-on-the-form-by-a Centrelink-employee-sitting-with-a-Yolnu-woman-on-the-lawn is worked to represent both money and child, and as a consequence relates the woman's two children to one hundred dollars of value in Australian currency. Words and numerals seem to effortlessly substitute children for money, effecting the reality of children as being self-evidently discrete and quantifiable in the same way as coins are self-evidently discrete and quantifiable.

Without interrogating this brief routine, nor accounting for the metaphysics affected by this enumeration (children and money as forms of discrete quantifiable entities), the act of the woman walking away can only be interpreted in a way that reproduces the separation of the moral and rational; either she morally rejects Centrelink and the Intervention, or she cannot understand the logic of money and finance. However, it may be that this mother and her two children are a single expression of one of the foundational relations constitutive of the Yolnu reality, yothu-yindi, and that attempts to fragment or undermine this are both morally and logically wrong. Hence, actively refusing this enumeration is the right and rational thing to do.

By recognising that the practices of enumeration in which numbers, worked as representations, effect particular metaphysics with particular subjects and particular objects, one can understand number as doing political work. Enumeration as representation helps to produce some realities and not others, and most of the time does so without question. Nevertheless, when we conceptualise numbers as neither ideal nor real objects existing independently of the practices taken to represent them but as objects crafted in this very practice, numbers are no longer guaranteed this secure self-evident existence. They can be outright refused, and they can also be crafted in slightly different ways, creating multiple types of number. This may sound odd--having multiple types of number--but this is exactly what has occurred in the continuing implementation of Income Management.

WHY YOU NEED A FOODCARD: NUMBER AS MULTIPLE

   It is a Thursday, a pay day. Today the Foodcards become active. The
   shop is busy. At the registers are three Centrelink staff, two
   Deloitte staff and the shop employees working on the registers. A
   woman hurries past holding her Foodcard, being led by Kate, a
   Deloitte staffer. Kate shows the shopper and a crowd around her the
   touch-screen computer kiosk where your balance is both displayed
   and sent to your Foodcard. The woman quickly walks to the counter
   to buy her goods. 'So who needs a Foodcard?' asks Kate in a loud
   voice.

   I grab a drink and wait in line, listening, watching and being a
   part of the confusion, occasional delight and frustration. In front
   of me is an older woman whom I had seen earlier at the shop
   accompanied by the Centrelink boss, the Government Business
   Manager, the homelands nurse, and a letter from Centrelink. Now she
   is at the counter, buying goods with her Foodcard. It is taking a
   while. A younger woman repeatedly adds more items to the pile of
   goods being registered for purchase. I hear the lady operating the
   register say an amount of money: 'three dollars eighty'. The
   younger woman leaves and returns with a bag of nuts. These are
   scanned and another amount is read out: 'one dollar twenty-five'
   she says, then suggests, 'apple?' The young lady fetches an orange.
   I realise that this little routine is trying to exhaust the money
   quarantined on the Foodcard.

   When this routine ends, the shop worker says, 'It's telling you,
   you need to pay.' Both buyer and seller are puzzled: isn't the idea
   of quarantining not to use money, not to 'pay'? The shop worker
   turns to the Centrelink employee who is sitting behind her and
   playing games on her laptop. 'Maybe she has spent too much?' the
   Centrelink employee responds. Kate comes over. 'She has one hundred
   dollars to spend...' Kate says and peers at the screen. 'For the
   chips and the baby bottle she has to pay cash,' she says, pointing
   to icons that must indicate this fact. 'The bottle costs two
   dollars forty and two fifty-five for the chips. Do you want to pay
   for them?' 'No.' The chips and the bottle are unregistered. 'You
   just don't know which items will be allowed on the Foodcard until
   they are scanned,' Kate says to the huddle of shop staff and the
   manager, Centrelink staff, myself and other shoppers who have
   gathered around this little lesson on Foodcards and Income
   Management.

Why do you need a Foodcard? To receive all of your payment? To eat healthily? Yes and no, and more. The Foodcard became necessary for participation in the emerging economy in which number, as representing money-value, became multiple. During the introduction of the Foodcard and the imposition of this multiplicity of number, I was in disbelief. I would come home from the shop full of stories such as the one above, needing to retell them in person, on the phone and in emails. Confusing and compounding my disbelief was that the Yolnu shoppers against whom, unlike me, this system discriminates were neither so shocked nor puzzled. The successful working of a Foodcard even provoked excitement and delight! How could I reconcile this?

Treating this as a puzzle, I refuse to adopted the position of an objective modern knower who might know the Yolnu as they don't know themselves: defeated and accepting, or ignorant of colonial injustice. In the preceding argument, I have shown how numbers are worked as representation, and how in recognising this we may also recognise other forms of number (and knowledge). However, while standing in the shop I was still holding onto number as accurate representations of money-value. Like many other people in shops across Australia, including in Ngunhili until this particular Thursday, I had experienced this to be the case. One would gather a collection of items from the shelves and take them to a checkout. There, the value of each item would be represented as a number, which in turn allowed the numbers to be summed into the total money-value of the collection of items. This was then paid for by money as cash or money on a bank card. Items represented as numbers, understood as money-value, were exchanged for money. This comfortable routine and understanding of numbers was disrupted when a number representing the money-value of a bag of nuts became different from a number representing the money-value of a bag of chips.

Customers purchasing items with a Foodcard had to present their Foodcard and insert it into the register before any item was scanned. Following this, the numeral representing the money-value of the nuts reduced the total on the Foodcard which was displayed on the screen. The numeral representing the chips was to be held separate to this, and only added together with numbers generated by other items on ALPA's non-nutritional list, such as baby bottles. These two sets of numbers were held apart, accumulated differently and paid for differently. That these two numbers cannot and should not be treated as one true number, I imagined to be the consequence of political interference. How has it come to be that numbers in contemporary economics and social sciences are understood as independent of politics? How has number achieved its apparent singular definite form which is accepted as the purest form of representation? Mary Poovey's work, A History of the Modern Fact, (22) provides an informative genealogy.

Poovey argues that numbers as numerical representation have come to epitomise the epistemological unit she calls the 'modern fact'. (23) The modern fact is the conceptual tool with which Poovey articulates the emergence of an understanding of knowledge that holds description and interpretation as independent tasks. However, as these two tasks can never be totally severed, the modern fact has developed to work both tasks while denying any relation between them. Numbers have come to epitomise this double but denied work of the modern fact, as Poovey says, 'because [numbers] have come to seem pre-interpretive or even somehow non-interpretive [descriptive success] at the same time as they have become the bedrock of systematic knowledge [interpretative success].' (24) Poovey's work traces the sustained effort to work numbers as separate from politics and morality in the emergence of the social sciences, especially economics. A potted version of Poovey's argument is helpful in familiarising ourselves with the troubling of single number caused by Income Management.

In the mid- to late-1700s, Adam Smith worked numbers as representing products that had been counted (the task of describing products sold), and also realising the good that was the liberal market economy (the task of developing a good society). (25) These numbers were understood and merited as both logical and moral. Within 100 years, however, Thomas Malthus was presenting the generation of numbers as removed from any moral work. He was vehemently challenged for relying on these now amoral objects of inquiry. Such criticism was averted, however, by John McCulloch, who initiated the tripartite separation of statistics, political science and political economy in the academy. The result was that statisticians, in the now solidifying social sciences of the nineteenth century, defended any questioning of the morality of their practice by arguing that they merely collected and combined numbers. (26) No more and no less. At this stage, numbers were worked as and understood to be completely cleaved from any interest, politics and morality. Number worked as representation, as an apolitical and amoral entity, dovetailed with a metaphysics which justified representation as having correspondence with an independent reality. The result was that the market economy took on the existence of an independent reality, and money became assumed to be self-evident as numbers.

In the above episodes, the troubling of number was not caused by the prohibited political interference in the domain of facts, but by the intense visibility of the efforts required to produce a new form of number, managed-money-value, while maintaining the former one, bank-money-value. The trick with which modern knowledge generates its power--denying its own messy practice while holding its constructed objects as always existing--was impossible to maintain. (27) The shop was overflowing with Centrelink employees, Deloitte employees, folders, computers, the touch-screen kiosk, Foodcards, confused shoppers, letters and forms. This extraordinary assemblage was there to ensure that the generation of numbers as managed-money-value was correct and stable. Politics (Centrelink and ALPA policy) and science (Deloitte and Foodcards) were inseparable. Depending on how this assemblage was arranged in practice determined which number was produced. Some objects made them more distinct--the insertion of a Foodcard or not--while others made them less distinct--the use of numerals. In some instances, you just had to wait and see if a baby bottle figured a routine as one producing number as managed-money-value or bank-money-value.

This fine-grained approach to the practices and routines of generating numbers reveals the complexity and potential multiplicity of practices and routines. A child picking up and stacking the red baskets in the shop saying 'one, two, three' is generating numbers with things, but these things are not for sale and do not generate numbers at the register. A mother who picks up the child and sits her on the counter does not expect the child to be scanned, weighed and represented as a number. Although in an interview with Centrelink while the child is at school, this child may indeed become represented as a number (as in the opening episode). Once we are willing to interrogate the generation of numbers, and resist the impulse to cut through these episodes severing facts and politics, number work and moral work, our analyses become more sensitive to situations such as the Intervention and Income Management. We begin to see these numbers as always managed, coming to life through particular arrangements of people, chips, words, cards and computers that enact number's inherent irreducible multiplicity.

WHERE IS THE MONEY? NUMBERS AS PERFORMANCE

   At the shop today, the Internet is down. It is five days after the
   Foodcard went online and it is now off line. The shop computers
   cannot connect to the Foodcard database, there are no updates on
   people's payments, and people cannot receive their payments.

   'So is the amount stored on the card, or is it just on the system
   when they put the card in?' I ask Kate, the Deloitte employee. 'It
   is on the card.' 'Really!' remarks Anthony, another Deloitte
   employee. 'I didn't know that.' Kate explains that the computer
   kiosk puts money onto the card, and that this is why people have to
   insert their card into the kiosk to receive money when it arrives.
   'So there's a lag between the system knowing and the card knowing
   about the money arriving,' says Anthony. Kate explains the routine.

   No card is given out with money on it. Upon receiving their
   Foodcard, Kate takes people to the kiosk and shows them how to
   insert it and where the balance appears. Kate says that developing
   the habit of going to the kiosk was central to training people for
   Income Management. 'While no second payments have come yet,' she
   says, 'people will have to remember to go to the kiosk to retrieve
   them when they do' (provided an Internet connection is present
   also). Finishing her description Kate concludes, 'It's all routine
   and habit. That's it.' 'Try telling a philosopher that numbers are
   just routine and habit!' I reply and we both smile.

Why did the loss of Internet connection make us think twice about the existence of managed money? Do numbers also disappear or get lost? Where are numbers? The final part of this paper suggests a solution to these questions by conceptualising numbers as embedded in the very routines in which they emerge and the ways of life they embody.

So where was the money? The managed money--that crucial object that affords Foodcard purchases--exists embryonically, distributed in the system (computers, databases, code, technicians and coffee mugs, one imagines). (28) Once the system coheres as the Foodcard database, shop, and kiosk all online together, a Foodcard may be inserted into the kiosk and attains money-value that is displayed as a numeral. Here, knowledge of the money entails becoming an embodiment of money. Anthony is surprised by this. Perhaps he was expecting the money to stay 'out there', existing on the system while what is transferred to the Foodcard via the kiosk is merely a representation of the money.

Where are the numbers, then, if not out there waiting as objective representations to communicate money-value? Helen Verran asked herself this question when puzzling over the goings-on in markets, classrooms and university labs in Yorubaland, Nigeria. Verran's analysis is a meticulous, patient and conscious development of a new approach to analysing number. According to Verran:

   Numbers are uncompleted, partial and distributed, located in their
   performance ... We could think of them as being there as complete
   objects only momentarily, ephemerally, in an accomplished
   performance. As enduring objects, they are located through matter
   and across space time ... Similarly, realities that numbers
   objectify are multiple, incomplete, infinitely partial, distributed
   and potential. (29)

How are numbers momentarily achieved in performance at the shop? Through, and as, the mundane routines described above in which two different performances come to generate two different numbers, managed-money-value and bank-money-value. However, although different, these two numbers are not in contradiction. One is not pure and the other is politically corrupted. They both exist as incomplete, distributed and partial in the shop, its shelves, products, policies, registers and customers. In many ways the numbers are the same--the materialities of the shop, for example--and continue to enact a political economy of bank money and a particular political order. In other ways they are different, and as such effect new political economies of managed money. In this new political economy, products become nutritious or non-nutritious and customers are either 'managed' or 'unmanaged'. Just as money is known through the cohering of a distributed existence in the system, numbers are also distributed and partial (in both senses of the term) and momentarily embodied as particular arrangements of subjects and objects.

The objectification number achieves, however, of money, products, and children, for example, is only ever a momentary clot. In answering the question, where are numbers? with a definition of numbers as distributed material routines of life, we make space for others ways of life that may perform number differently again, or ones that may be enough without numbers at all. Numbers, and facts, are no longer accessible to modern research in being outside politics and morality. Rather, they participate in moral questions about what are good and desirable ways of living. How these are arranged in our research and our accounts of it matters. According to Verran:

   This choosing to arrange things for a going-on together, or a
   goingon only in this way, is a politics. It is a politics built
   either on trust and commitment to a community here and now, or a
   politics of imposition, a commitment to a knowledge community that
   is not here and now. This latter strategy amounts to an intimate
   form of colonizing. (30)

Povinelli's generative impasse is an important tool in forming one's commitment to establishing knowledge in research. As an analytic tool, generative impasses provide a way to begin engaging with the limits of Western ways of knowing and living. More importantly, however, as Povenelli points out, analysis is only interesting if one accepts these impasses as generative and learns from the emerging ways of life that they perform.

CONCLUSION

This paper has addressed the problem of engaging meaningfully with a generative impasse in which rationality and morality become enmeshed. It has done so by telling three episodes of Income Management and, through them, making three moves in re-conceptualising the hardest case of the modern fact: numbers. Numbers are found to be produced and worked as objective representations of a singular real. They appear to contribute effortlessly to both description and interpretation, while holding these two tasks separate. Income Management troubles this single number representing money-value by crafting multiple numbers: managed-money-value and bank-money-value. This troubling makes visible the routines in which numbers are produced, which informs the analytic work of re-conceptualising number. The understanding of numbers developed here as embodied routines, distributed and partial recognises all numbers as managed in practice and managing in the reproduction of everyday life. This re-conceptualisation is important, as it allows room for new and multiple accounts of numbers, money and markets in remote Indigenous Australia. It locates research and analysis in the situations of their engagement, and provides a framework with which the mainstream Western academy can engage more meaningfully with those who live at and beyond its limits.

Christian Clark

School of Philosophy, Anthropology and Social Inquiry

ENDNOTES

(1) I wish to thank the people in community where my fieldwork was undertaken, the staff at the Arnhemland Progress Association, Deloitte and Northern Land Council for their support during the fieldwork, Helen Verran for her conversation and comments during the writing of this paper and Lara Thurlow for her support and help in editing. I would also like to thank the three anonymous reviewers for their engagements and helpful comments.

(2) Ngunhili means 'at that place not visible to speaker' in the Gapupuyngu, the language of the traditional owners of the land on which the community was established. I used it here as name and figure for the community; it is clearly not an English nor Western word, but has been made and written in a way familiar to Western eyes and tongues.

(3) Yolnu is the term used by Indigenous people of North East Arnhem Land to refer to themselves and often other Australian Indigenous peoples.

(4) This mysticism is not unique to rationalist traditions. Romantic traditions, common in anthropology, literature and public debate, equally make the seemingly contradictory move of grounding knowledge of an 'other' in a mysticism.

(5) Desert Knowledge Australia, remote FOCUS: Revitalising Remote Australia, Prospectus, November 2008, www.desertknowledge.com.au/dka/index. cfm?fuseaction=remoteFocus

(6) Patricia Anderson and Rex Wild, 'Ampe Akelyernemane Meke Mekarle--Little Children Are Sacred, Report of the Northern Territory Board of Inquiry into the Protection of Aboriginal Children from Sexual Abuse', Northern Territory Government, Darwin, 2007.

(7) Rex Wild, 'Unforeseen consequences', in John Altman and Melinda Hinkson (eds), Coercive Reconciliation: Stabilise, Normalise, Exit Aboriginal Australia, Arena Publications, Melbourne, 2007, 111-121.

(8) John Altman and Melinda Hinkson (eds), Coercive Reconciliation: Stabilise, Normalise, Exit Aboriginal Australia, Arena Publications, Melbourne, 2007, 1.

(9) Yu, Duncan, and Gray, 'Report of the NTER Review Board October 2008'.

(10) Annemarie Mol, The Body Multiple: Ontology in Medical Practice, Duke University Press, Durham, 2002, 5.

(11) Bruno Latour, Politics of Nature: How to Bring the Sciences into Democracy, Harvard University Press, Cambridge, Mass., 2004.

(12) Helen Verran, Science and an African Logic, University of Chicago Press, Chicago, 2001, 158.

(13) Elizabeth Povinelli, The Cunning of Recognition: Indigenous Alterities and the Making of Australian Multiculturalism, Duke University Press, Durham and London, 2002, 9-10.

(14) Povinelli. 9-10.

(15) Paul Toohey, Last Drinks: The Impact of the Northern Territory Intervention, Black Inc., Melbourne, 2008, 74-5.

(16) Northern Territory Anti-Discrimination Commission, 'Message from the Commissioner', Newsletter of the Northern Territory Anti-Discrimination Commission, June 2008; John Gardiner-Garden, Coral Dow and Michael Klapdor, 'Budget 2009-10: Indigenous affairs', Parliament of Australia: Parliamentary Library, 2009, www.aph.gov.au/library/Pubs/RP/ BudgetReview2009-10/IndigenousAffairs.htm, accessed 22 September 2009; Australian Broadcasting Corporation, 'Indigenous intervention discriminatory', www.abc.net.au/news/stories/2009/08/27/2668915.htm, accessed 22 September 2009.

(17) Commonwealth of Australia, 'Top End Women's Legal Service', Northern Territory Emergency Response Review, 2008, www.nterreview.gov.au/subs/ nter_review_report/181_Top_End_Womens_Legal_Service.htm

(18) Arnhemland Progress Association, 'ALPA Arnhemland Progress Association NT Australia', www.alpa.asn.au/updates/foodcard.php

(19) Ian Hacking, Representing and Intervening: Introductory Topics in The Philosophy of Natural Science, Cambridge University Press, Cambridge [Cambridgeshire], 1983, 133.

(20) Hacking, 139.

(21) Hacking, 142.

(22) Mary Poovey, A History of the Modern Fact: Problems of Knowledge in the Sciences of Wealth and Society, University of Chicago Press, Chicago, 1998.

(23) Poovey, xii.

(24) Poovey, xii.

(25) Poovey, 216.

(26) Poovey, 301-4.

(27) Bruno Latour, We Have Never Been Modern, Harvard University Press, Cambridge, Mass., 1993, 43.

(28) Adrian Mackenzie, 'These things called systems: Collective imaginings and infrastructural software', Social Studies of Science, vol. 33, no. 3, 2003, 365-387.

(29) Verran, 106-7.

(30) Verran, 117-8.