Human Rights Indicators and the Sovereignty of Technique

Indicators are increasingly used in international human rights monitoring, and time, expertise and resources are being devoted in ever-growing quantities to the production of more apparently powerful and sophisticated ways to objectively measure human rights performance. However, there is a certain level of resistance and scepticism to the statistical measurement of human rights on the part of many practitioners and advocates, who argue that it is reductionist and disruptive to their work. This article uses the writing of Michael Oakeshott as a lens through which to examine the shift towards indicators and argues that it is a project that is strongly characterized by rationalism: a desire for certainty, uniformity and clarity that neglects the experiential, tacit, and conversational. This not only provides a method for analysing the dangers present in the phenomenon but also explains why the reliance on indicators and other measurement methods seems destined to grow despite the reservations held by practitioners and scholars alike.


Introduction
Mary Robinson once envisioned a 'science of human dignity' based on the use of statistics in monitoring human rights. 1 For good or ill, the quest to construct this science is well underway. The monitoring of human rights performance is increasingly dominated by a culture of indicators, benchmarks and statistical measurement. Perceived as a way to more accurately assess compliance, and to better elaborate the content of rights, indicators in particular have come to occupy a central role in the agenda of the United Nations (UN) treaty-based mechanisms. At the same time, there is a trend among human rights advocates to devote ever-increasing time and energy towards 386 EJIL 27 (2016), 385-408 the creation of more powerful, useful and specifically tailored statistical methods for measuring human rights performance.
This effort marks a significant shift in emphasis away from what might be called the 'traditional' approach to human rights monitoring, which was largely (though by no means entirely) carried out as a discursive or narrative-based process. States parties would describe their efforts made towards the protection and promotion of human rights, and treaty body members would consider those efforts based on their own expertise and judgment alongside reports from local advocates and non-governmental organizations (NGOs). The better approach is increasingly seen as being a process resembling an audit, certifying that relevant indicators are produced -and produced in the correct disaggregated fashion -and assessing the accuracy of the data.
It would be an exaggeration to say that this shift signifies a rift in 'the human rights community', if such a thing exists. Most of those who support an increased role for indicators see them as simply another tool in the arsenal of human rights monitoring to go alongside existing mechanisms. Yet there is a concern that indicators are transforming 'from a tool to a paradigm' -that instead of being complementary to traditional methods, they could come to dominate or supplant them. 2 Indeed, there is a sense that this is already taking place, particularly given the developments within the Office of the High Commissioner for Human Rights (OHCHR) and its efforts to systematize and standardize the uses of indicators within the UN treaty body system. Likewise, developments in academia in the field of the statistical measurement of human rights have proliferated with, undoubtedly, an attendant opportunity cost regarding other varieties of research. Some scholars have expressed concern with this apparent shift towards audit-based procedures -a shift that marginalizes questions of human judgment and, instead, focuses on methods of verifying whether and how data is produced.
This article argues that this shift can be readily attributed to a strong rationalist propensity, as Michael Oakeshott would have described it, present within the move towards indicators. That is, the push for ever-more accurate and pervasive statistical measurement is strongly characterized by the application of technique -by ever-greater centralization, uniformity and certainty, which means that, as with all rationalist projects, it tends to grow increasingly rationalist over time, and its rationalism comes to dominate and crowd out other methods of thinking.
The move to indicators is ripe, then, for analysis from an Oakeshottian perspective. The analysis presented here not only encompasses existing concerns about the domination of human rights monitoring and discourse by indicators -the dangers of unanticipated consequences and the chilling effects that indicators have on conversation and dialogue -but also provides an explanation for why the project continues to grow; why it appears to represent not just a supplementary tool to traditional human rights monitoring but also the transformation of a tool into a paradigm that displaces other perspectives. The article begins by describing the shift towards the use of indicators in

The Human Rights Indicator Project
It is difficult to think of areas in the public sphere in the developed world, or in international governance in general, that have not come to be characterized by the use of purportedly objective, mostly numerical, measures indicating achievement or performance and the resultant league tables and rankings. Some prominent examples listed in a recent collection on indicators 3 are the Programme for International Student Assessment, which measures the educational achievements of 15-year-old students across the Organisation for Economic Co-operation and Development; 4 the state fragility index used by the US Agency for International Development; 5 and the yardsticks for the quality of national laws regarding corporate bankruptcy used by the World Bank and the European Bank for Construction and Development when making decisions on lending. 6 In the domestic sphere, examples from the literature are the US News and World Report's ranking of US law schools; 7 performance indicators used by the Audit Commission to produce national league tables for local authorities in England and Wales across various national indicators; 8 and the Research Assessment Exercise (now Research Evaluation Framework) used within the audit of British higher education institutions. 9 The field of international human rights is no exception to the general trend towards the use of indicators and quantitative measurement. Indeed, in recent years, indicators have become something of a fad among advocates of economic, social and cultural rights and are commonly presented as a remedy to a perceived sense of those rights' marginalization. 10 The shift to indicators is often seen as having come into being alongside a broader move towards an improved conceptualization of economic, social and cultural rights over the past several decades. 11 This is not to say, however,

388
EJIL 27 (2016), 385-408 that the realization and monitoring of civil and political rights has fallen outside of its purview.
Human rights indicators and benchmarks have been in use since at least the early 1990s, when the Special Rapporteur on the Realization of Economic, Social and Cultural Rights recommended considering how indicators might be used in measuring progression towards the realization of such rights. 12 Within the UN treaty body system, this largely took place in an ad hoc fashion as treaty bodies developed their working methods and interpretations of treaty provisions. For instance, in 1999, the Committee on Economic, Social and Cultural Rights (CESCR) issued its General Comment No. 13 on the Right to Education, in which it suggested that there was a minimum obligation for state parties to create a national education strategy allowing monitoring through the use of indicators and benchmarks. 13 At the same time, however, there have been efforts to make the production of indicators and their usage more systematic and widespread. A reader familiar with the literature on the quantitative measurement of human rights performance will be aware of a number of such projects. These include the SERF Index developed by Sakiko Fukuda-Parr and colleagues, 14 Todd Landman's concepts of rights in principle, practice and policy and their subdivision into indicators 15 and the project to develop a specific suite of indicators for the right to health. 16 By far the most significant such effort, however, has been the project undertaken by the UN's OHCHR, initially at the behest of the chairs of the treaty bodies themselves, to systematize the production and usage of human rights indicators. Beginning in 2005 and culminating in the 2012 Human Rights Indicators: A Guide to Measurement and Implementation, this project has resulted in a conceptually unified approach that aims to produce a core set of 'universally relevant' human rights indicators and a generally applicable method for generating them. 17 This document has three main pillars. First, the preference is for the quantitative over the qualitative: '[F]act based or objective indicators, in contrast with judgement-based or subjective indicators, are verifiable and can be easier to interpret when comparing the human rights situation in a country over time and across populations.' 18 By the same token, qualitative indicators based on expert judgments are considered to be unreliable, unrepresentative, lacking in transparency and generally of a 'limited purpose'. 19 The guide subdivides indicators into four categories labelled A, B, C and D and considers A -indicators that are quantitative and objective -to be the most useful. 20 Even where intrinsically subjective phenomena are being measured, the preference is for quantitative data that might emerge through statistically representative surveys (one example being the 'percentage of individuals who feel safe walking alone at night'). 21 Where indicators based on expert judgment are to be used, they are to be coded in the first place in order to convert them into numbers or symbols that can be tabulated and counted. 22 Second, in order to 'concretize' the content of each right and make explicit the link between the normative standards of a given right and its indicators, it was decided that attributes -usually around four in number -ought to be identified within each right. The aim was to provide a way to break down the many different rights in a way that facilitated the easy selection and measurement of indicators. For the right to life, which is the main example given, these attributes were 'arbitrary deprivation of life', 'disappearances of individuals', 'health and nutrition', and 'death penalty'. 23 The same approach, it was suggested, should be taken with all rights, drawing from all of the nine international human rights treaties to determine a handful of attributes that could then be used as the basis on which indicators might be selected. 24 Third, it was made clear that the framework ought to revolve around Avedis Donabedian's model of structure-process-outcome, 25 which had originally been created in order to evaluate the quality of care in health services. The essential aim was to create a set of structural, process and outcome indicators for each attribute of each right, which would in effect measure a given state's commitment, effort and results respectively. This process would help 'unpackage the narrative on the normative content of a right', 'demystify the notion of human rights and take the human rights discourse beyond the confines of legal and justice sector discussions' and so on. 26 The sample indicators on the right not to be subjected to torture or cruel, inhuman or degrading treatment serve as an illustration of the OHCHR's aim. 27 As usual, the right is divided into a handful of separate attributes: 'Physical and mental integrity of detained or imprisoned persons', '[c]onditions of detention', '[u]se of force by law enforcement officials outside detention' and '[c]ommunity and domestic violence'. Each of these has a common set of structural indicators, which include 'international human rights treaties relevant to the right … ratified by the State', '[d]ate of entry into force and coverage of the right not to be tortured in the constitution or other forms of superior law' and '[d]ate of entry into force and coverage of domestic laws for implementing the right not to be tortured, including code of conduct on medical trials 20 Ibid., at 19. 21

390
EJIL 27 (2016), 385-408 and scientific experimentation on human beings'. The indicators then become subdivided into further structural, process and outcome indicators. Taking one of the attributes, '[c]ommunity and domestic violence', a structural indicator is '[d]ate of entry into force and coverage of specific legislation on community and domestic violence'; a process indicator is '[p]roportion of women reporting forms of violence … against themselves or their children initiating legal action or seeking help from police or counselling centres' and an outcome indicator is '[i]ncidence and prevalence of deaths and crimes related to community and domestic violence … in the reporting period'. For this single right, across the four attributes, there are 10 structural indicators, 19 process indicators and nine outcome indicators provided. This is broadly illustrative of the 14 rights that are examined in the guide.
The OHCHR's approach has now begun to be adopted across the UN treaty bodies, and the production of indicators and the collection and disaggregation of data with indicators in mind can only be expected to draw increasing time, attention and resources from governments, practitioners and academics alike. At the same time, however, there is a certain level of resistance against the move towards indicators, not least from some members of the treaty bodies themselves, who have been described as 'lukewarm'. 28 The quantitative measurement of human rights has been viewed as being problematic almost since its inception, 29 and just as the indicator project has gathered momentum, so have the concerns expressed about it. Indeed, it is rare to find an academic article written about the subject that has not acknowledged severe problems with, at the very least, the reliability of statistics, data collection and disaggregation. 30 More seriously, some scholars have seen in the project a shift in focus towards auditing: a change in emphasis from judgment-based decision making to an exercise in verification and checking -in other words, a reconceptualization of the role of the treaty bodies away from the judgment-based, subjective assessments of state parties' performance and specific, contextual recommendations towards the objective evaluation of data collection methods and the verification of outcomes. This brings with it the concern that monitoring through indicators ignores the complexity of what human rights represent -that no level of quantitative measurement or benchmarking can displace the 'pesky, irreducible core of human judgment' that must always be the final arbiter of which policy is implemented, how resources are allocated and how choices are prioritized. 31 That is to say, both the nature of human rights monitoring and the nature of human rights practice is highly dependent on context and, hence, human judgment and expertise, and this cannot be replaced simply with the use of more and better indicators and benchmarks.
What explains the rise in the use of indicators, despite these concerns? And why should proponents of human rights view it with such caution? This article argues that all of these questions can be answered by considering the move towards indicators as a project characterized by a propensity towards rationalism.

Oakeshott and Rationalism
Oakeshott defined rationalism by reference to a character -the rationalist. He describes this person as standing, at all times, for independent thought, free from obligation to any authority except for reason: 'He believes … in the open mind, the mind free from prejudice and its relic, habit. He believes that the unhindered human "reason" … is an infallible guide in political activity.' 32 This makes him, on the one hand, partially a sceptic because he never hesitates to subject any opinion, habit or belief to the power of his reason. He has no capacity or willingness to accept the 'mysteries and uncertainties of experience' or the traditional knowledge of his society at face value. Yet, on the other hand, he is also an optimist: he has absolute faith in the power of reason and rational argument to solve mankind's problems through the pursuit of universally valid answers. 33 His political activity consists in bringing the 'social, political, legal and institutional inheritance of his society before the tribunal of his intellect' -a project of innovation and construction; of solving problems and scrubbing clear the tabula rasa of the 'irrational scribblings of tradition-ridden ancestors'. 34 Nothing is to be accepted merely because it exists or has existed for generations but only because it passes the test of reason.
Oakeshott's rationalism has a number of different characteristics. First, it views all problems as being resolvable; it transforms the sphere of political action from the open-ended and conversational to 'something more like an argument, which demands a conclusion'. 35 Politics is that of 'felt need' -continuous resolution of crises through the application of reason. 36 It is not a project of making do or the careful amendment of traditional rules and practices but, rather, the use of rationality unclouded by 'the fumes of tradition' to construct new rules, new inventions and new solutions. 37 It is the politics of the blank slate: a preference at all times for the well-reasoned solution based on abstract rationality over the messy but tried and tested.
Second, rationalism has within it what Oakeshott calls a 'doctrine of human knowledge', the result of which (or perhaps the cause) is a preoccupation with certainty. 38 He examines this through an extended musing on knowledge and how it can be sorted 32 Oakeshott, 'Rationalism in Politics', in M. Oakeshott (ed.), Rationalism in Politics (1962) 1, at 4. 33 Ibid., at 1-3. 34 Ibid., at 5. 35 Candreva, The Enemies of Perfection (2005), at 28. 36 Oakeshott, 'Rationalism in Politics', supra note 32, at 5. 37 Ibid., at 4-7. 38 Ibid., at 7-13. into two separate categories that are akin to the ancient division between technē and epistēmē. That is, knowledge is generally divisible into two sorts: practical knowledge (knowledge that is generated by and contingent on practice) versus technical knowledge or technique (knowledge that is formulated into rules). The difference is that which lies between the craft of a chef and the recipes contained in a cookery book or between the skill of driving a car and the rules contained in the highway code. Nobody supposes that the knowledge a good cook has is confined to what can be contained in a book; he also has what Michael Polanyi would have referred to as 'tacit knowledge' -what one has learned through practice (how different tastes combine, how much salt to add to a particular dish, how to replace ingredients with others and many unspoken rules of thumb). 39 Essentially, all human activities -driving, cooking, painting, writing poetry, science -are comprised, according to Oakeshott, of both of these types of knowledge: the formal rules that must generally be learned and that compromise technique, and the informal, experiential, tacit understanding of the craftsman or practitioner. Moreover, the technical and practical can never be separated or used as a replacement for each other. Thus, a pianist acquires not only technique (how to read music and so forth) but also artistry; a chess-player learns style and insight a well as what the potential moves are; and a scientist acquires intuition about profitable avenues to explore as well as the theory of his field. 40 The rationalist, according to Oakeshott, believes that all real knowledge is technical knowledge and, therefore, that practical knowledge is not really knowledge at all -it is rather a 'nescience, which would be negligible if it were not positively mischievous'. 41 While practical knowledge seems imprecise and uncertain -a matter of opinion -and therefore not to be trusted, technical knowledge is clear and absolute. The rationalist therefore holds reason sovereign; he believes in the 'sovereignty of technique'. Concomitantly, he believes in certitude: technical knowledge gives the appearance of being certain, in that it is self-contained and not predicated on prior knowledge, prejudice, preconceptions -or traditions. In its certainty, it is complete and providing of 'order and distinctness'. 42 It is the reduction of the 'tangle and variety of experience' to a 'set of principles' that can be assessed upon rational grounds. 43 Third, as a result of both its capacity to provide final resolutions and its certainty, and likewise its scepticism of tradition and its suspicion of practical knowledge, rationalism tends to have as one of its goals the purging of habit and prior prejudice. Since technical knowledge has the appearance of being self-contained, and since the rationalist believes in the supremacy of technical knowledge, other forms of knowledge can be ignored. Indeed, this is desirable to avoid being governed by 'nescience'. In other words, the rationalist expects and desires certainty, believes that technical knowledge is the only kind of knowledge that can provide it and, hence, requires the expurgation of anything that may hinder it -like the desire of Voltaire to have good laws by burning all existing laws and starting afresh. 44 Thus, the rationalist, as well as viewing problems as being permanently resolvable, also believes in the certainty of technical knowledge to provide such resolutions and, as a result, believes in the scrubbing out of what comes before it.
Readers sometimes come to the conclusion that Oakeshott is developing something of a straw man in his portrayal of the rationalist 45 (or, worse, is simply using the figure as a kind of stand-in for socialists, progressives or liberals). 46 Others view his theories as being outdated -a reaction against the world in which Oakeshott lived, with its recent history of domineering ideologies and its reformative, progressive postwar fervour. 47 Others still construe his position as being nihilistic -a refutation of reason itself. 48 However, Oakeshott was keen to emphasize that rationalism was not merely a phenomenon of the era in which he was living and writing. This is because, far from referring specifically to a doctrine or a philosophy, Oakeshott's description of the rationalist is really a personification of a certain conception of knowledge. 49 His essay is not to be understood, then, as a polemic against a certain class of people or ideas but, rather, against a certain method of thinking. It is a propensity rather than a person. Moreover, it is a propensity present in everyone and everything. In politics, rationalism has 'come to colour the ideas, not merely of one, but of all political persuasions, and to flow over every party line', and it is discernable in all fields of human activity. 50 At the same time, however, it is nowhere absolute or complete, and the propensity can only ever be partially embraced -the rationalist is a non-existent extreme. Or, rather, 'the Rationalist is no-one, and yet he is everyone'. 51 Oakeshott's critique is not of a person, a political persuasion or a movement but, rather, of a tendency that we all share to some degree or other. Oakeshott's description of the rationalist and rationalism provides us with a compelling account for the international human rights system's move towards the use of indicators, as the next section makes clear.

The Propensity to Rationalism in the Indicator Project
The rationalist character of the use of indicators to measure human rights performance is uncontroversial, if we use the word 'rational' in its ordinary meaning. Indeed, it seems unlikely that any of those engaged in the advocacy or production of indicators would disagree. The language of the OHCHR is of systematization, objectivity and science, and a prominent recent volume on the measurement of human rights is almost explicit 394 EJIL 27 (2016), 385-408 in stating the case for a neutral, empirical approach that sets to one side foundational, philosophical or normative concerns. 52 Using the term in its dictionary meaning, the approach is self-consciously rational or purportedly so.
Oakeshott's rationalism, however, has certain important characteristics. He did not simply mean by 'rationalism' to refer to anything that purported to be objective or scientific. He was referring to a more specific mode of thinking that was constructed around the use of reason unfiltered by tradition or habit to provide resolutions to problems and 'felt needs'; the preoccupation with certainty, finality and uniformity that comes from the sovereignty of technique; and the desire to expurgate prior prejudice and override the practical, traditional and experiential. It is the application of abstract and universal technical knowledge to any and all problems -or the use of premeditated principles to arrange a political community. 53 Put this way, the move towards indicators clearly has something of the character of Oakeshottian rationalism about it.
In the first instance, and most prosaically, the use of statistical indicators can certainly be conceptualized as the application of technical knowledge to a single, big, perceived problem -what Landman and Edzia Carvalho refer to as the 'large and variable gap between the expectations for human dignity outlined through human rights standards and the reality of the precariousness of those rights as they are variously enjoyed around the world'. 54 That is, there is one large and pressing issue in the field of international human rights law -the disparity between agreed obligations and actual performance -and this can be resolved, it is argued, by the use of objective measurement. As Judith Welling puts it, in a passage that neatly summarizes this view: By improving the accuracy and comprehensiveness of data relating to the status quo, international indicators assist states parties in better understanding the current situation ... By providing feedback on the outcomes of state policy, international indicators can inform decision making and suggest areas for future policy change at the governmental level … Illuminating the current conditions will thus help state members to better understand the consequences of their policy decisions as well as how they might improve their ability to satisfy ESCR commitments. 55 That is, objective measures supposedly allow for making objective judgments about progress, which can only be to the good since it will provide information on what has been achieved and what remains to be achieved and, thus, equip relevant parties with the necessary knowledge to make progress. A set of other advantages logically follow from this given and are summarized in Landman and Carvalho's book as allowing contextual description and documentation, classifying and categorization, monitoring of compliance, mapping and pattern recognition, secondary analysis and advocacy. 56 Such advantages would all have the general ancillary benefit of making the work of the monitoring bodies more efficient and streamlined. 57 Many of these benefits are 52 Landman and Carvalho, Measuring Human Rights (2010), at 1. 53 58 and it is not unusual to see advocated in the literature a fostering of a 'culture of statistics' within the international human rights system in general, with many of these benefits in mind. 59 Even where scepticism is raised about the possibility of measuring human rights performance empirically, and the underlying assumptions behind the production of indicators, there usually remains a sense that what are needed are simply better indicators used in different ways.
The usefulness of indicators in this respect, however, is even greater from a purposive perspective. Indicators 'link the conceptual discussion about human rights compliance to implementation practices'. 60 They do not merely measure human rights compliance in the abstract; they also instigate movement in pre-determined directions and supply ready-made policy goals. Implicit in any indicator is a final destination or, at the very least, a direction of travel. The indicator 'proportion of women reporting forms of violence … against themselves or their children initiating legal action or seeking help from police or counselling centres' has within it an implied goal of improving this proportion and the ultimate target of increasing it to 100 per cent. However, it also readily lends itself to benchmarks -increasing the proportion from 5 per cent to 10 per cent and so on -and, indeed, the OHCHR guide makes great play of benchmarking not merely in human rights monitoring but also in the performance monitoring of development programs and government policy interventions as a whole. The use of human rights indicators in governance, then, is not merely a method of measuring compliance but also a measure of building certain human rights targets and goals -and, indeed, norms -into government policy. For human rights advocates, this is clearly something that is intrinsically desirable.
At the same time, indicators imply the existence of ideals. 61 The World Bank makes their normative character explicit: 'In a sense, they are also communicative instruments.' 62 They are not merely data but also statements of what is desirable, which means that they can express values: the adoption of a set of universal indicators 'would be a high-profile reinforcement of the claim that those human rights set out in international law are indeed universal'. 63 This gives indicators a certain political usefulness. Human rights indicators, especially if standardized and universalized, give substance to the normative position that human rights are available to all individuals on the basis of interrelatedness, interdependence and indivisibility. 64 But, of course, individual indicators also have a normative function in expressing not only a purportedly empirical measure of compliance or 396 EJIL 27 (2016), 385-408 performance regarding a certain right but also a vision of what that right ought to mean and what its functions ought to be. The indicator 'proportion of women reporting forms of violence … against themselves or their children initiating legal action or seeking help from police or counselling centres', for instance, makes a clear statement about the content of the right being measured. It makes explicit that this right is not merely concerned with the traditional focus on the judiciary and the criminal justice system but also includes the prevention or punishment of domestic violence. It also makes explicit, through a structural indicator, that specific legislation on community and domestic violence is a desirable mechanism for doing so. Irrespective of whether this interpretation is to be welcomed, it illustrates the power of an indicator to make statements about the width and scope of a given right. Of course, this can be done as much through state-specific indicators as it is through universal ones, though it may be noted that the setting of 'attributes' by the OHCHR has much of this character. In defining the attributes of a given right, the OHCHR is not merely expressing empirically what it considers the nature of that right to be but is also implicitly including and excluding areas of public and private life from its scope.
It is apt, then, to think of the indicator project as being a response to a set of 'felt needs' -whether they are for more information and measurement per se, for the goal of human rights 'mainstreaming' in policy and budgeting generally, for clarification and elaboration on the nature and content of rights or, indeed, for expressing normative values. In turn, these serve the larger felt need of bringing states' behaviour in line with their obligations and of expressing certain standards and ideals. Moreover, it is a response that attaches great importance to the rationalist principles of certainty and uniformity.
Of course, no UN body has taken the view that a universal standardized set of indicators that applies always and everywhere is required. While the OHCHR's stated position is that the 'core content' of human rights ought to be monitored through the use of universal indicators, its desire is not for a complete and concrete suite of indicators for every single human right in every single state party to every treaty. 65 Yet while there may never be a complete set of universal indicators, it is nonetheless the case that what has been advocated is the universal use of indicators and a universal framework guiding their use. Uniformity and certainty, in other words, manifest themselves not at the level of direct implementation but, rather, at the level of conceptualization -it is in the overarching structure of the international human rights system in which uniformity and certainty appear.
The desire for certainty most clearly finds its expression in the emphasis on objectivity. As Sally Engle Merry puts it, much of the power of indicators lies in their capacity to 'convert complicated contextually variable phenomena into unambiguous, clear, and impersonal measures'. 66 The objective transforms the nebulous into the concrete.
Much of the indicator literature advances the argument that all forms of data -qualitative and quantitative, subjective and objective -are in some respect mutually reinforcing, but there is generally an implied or express preference for the quantitative and the objective. While qualitative and quantitative data may be viewed as complementary, the qualitative and the subjective is often represented as being in some sense inferior in usefulness, primarily because it is lacking in certainty and, thus, in need of reformulating in such a way that it can be numerically measured -repackaged so the subjective becomes less so. The OHCHR's guide, with its four quadrants of indicators ranked A-D, puts this most starkly, but a similar implicit hierarchy is present elsewhere -for instance, in Gauthier de Beco's 2008 article, which seems to place household perception data below socio-economic data and events-based data in terms of its usefulness for monitoring human rights performance and places expert judgments lowest of all, 67 or Craig Mokhiber's desire to bring subjective measures within the rubric of data such as opinion polls and other more 'objective' measures of public confidence. 68 As well as being more accurate and clear, a crucial element cited as an advantage of using quantitative indicators is that they provide consistency. Unlike the subjective opinion of an expert, an indicator allows close tracking of performance over time, permitting the accurate assessment of improvement or failure. 69 Yet the desire for uniformity and certainty also manifest themselves as broader guiding principles, both within the initial impetus towards the wider use of indicators and their continuing expansion. Obviously, the OHCHR's efforts are straightforwardly an attempt to systematize and universalize the production and use of indicators -if not to achieve uniformity at the level of individual indicators used, then to achieve it in the form of a 'common approach'. 70 However, there is also an aspect of the drive towards indicators that seeks not merely to measure but also to map; if not to definitively set out in concrete terms where the limits of each right lie and what their respective characteristics are, then at least to rationalize the process by which this could be determined.

The Consequences of the Rationalist Propensity
The indicator project, then, can be aptly described as possessing a strong propensity towards rationalism. This analysis now turns to the potential consequent pernicious effects that, indeed, have been identified by scholars who are sceptical of the move to indicators. The first of these -the production of unintended consequences -is obvious and something that Oakeshott dealt with briefly. This is not the concern that rationalist endeavours will necessarily result in bad outcomes (although this certainly may 67  be the case 71 ), but it is rather that, being weighted towards only half of the available sphere of knowledge (the technical), rationalist projects can only at best be half right. Therefore, they must result in unforeseen effects. While Oakeshott did not use this language, he was clearly alluding to something similar to Robert Merton's 'unanticipated consequences of purposive social action' 72 -a theme that scholars have subsequently teased out of Oakeshott's work. 73 That is to say, an approach that is at best half right will also be at best half wrong and often in such a way that its proponents cannot recognize.
Unanticipated consequences must manifest themselves in the indicator project because both the measured and the users of measures have the tendency to change their behaviour as a result of the measurement concerned. Measures 'create social worlds' by causing people to think and act differently. 74 Michael Power, in his work on auditing, divided such reactions into decoupling and colonization. In the first instance, there is a strong incentive for the subjects of an audit to attempt to render the process 'ceremonial' -to produce comfort in the auditing body through ritualized compliance and the production of 'auditable form' rather than actual substance. 75 In other words, states may simply 'buffer away' the monitoring process by going through an ineffectual, but apparently exhaustive, set of checking and measurement, issuing of technical guidance, production of measures and metrics and so forth. 76 As Power argues, while such efforts can never be completely successful, they shift the focus away from the proper moral concern (in the case of human rights, the protection of individual human beings) to something altogether more abstract and ultimately often meaningless: the creation of auditable outputs to satisfy external monitors as an end in itself. 77 To this might be added the opportunity cost of diverting resources and time towards the creation of auditable performance (indicators, standards, measures and associated data collection and disaggregation) as opposed to the actual protection of individual citizens' human rights.
In the second approach, precisely the opposite concern is that the values and practices of audit permeate an organization -or a state -to such an extent that it creates new mentalities, new strategies and new goals that interact in hitherto unknowable or unpredictable ways. Actors change their behaviour and activities in the name of these new values, and they do so in complex ways that may result in unintended and undesirable outcomes. Power uses the example of the Research Assessment Exercise in United Kingdom (UK) universities, which caused academics to prefer publishing in journals to books and to prefer not to spend time doing activities that were not measured, such as editing books and reviewing publications. 78 While these effects are not catastrophic for higher education or research in the UK, they are certainly different from what was desired or intended. We can expect similarly unpredictable effects to take place wherever an audit takes place, with the danger not that they are necessarily negative but, rather, that they are unknown and unknowable ex ante.
As Power puts it, decoupling and compartmentalization are inevitable because individuals are 'infinitely more complex and adaptable than normalizing attempts to measure and control them', 79 but at the same time, colonization through auditing must also always be expected, because motivations will always tend to become aligned with targets. At its worst, this produces an indicator culture that simply creates a gameplaying mentality for the audited and 'reverse effects', in which the production of auditable performance actually undermines the initial aim. 80 Of greater concern to Oakeshott, however, was that the twin desires of the rationalist -for, on the one hand, uniformity and certainty and, on the other, the application of principle distilled from experience -would diminish and displace discourse; that they would close conversation through the application of reason to finding the best or correct course. And, indeed, this danger is immanent in the human rights indicator project. In the first place, if the content of rights has been determined through attributes, and delineated through specific indicators, then there must be a chilling and finalizing effect on conversation and dialogue. For if a set of indicators and attributes defines a given right's content, then what need is there to discuss alternative visions of that right? There are merely numerical measures by which to assess performance. This, indeed, was foremost among Oakeshott's concerns. Whereas unintended consequences are almost mentioned as an aside, he was particularly anxious that the aim of rationalist projects was to provide finality and certainty rather than a continual making do.
The capacity for indicators to reduce and close off conversation has been widely acknowledged. Ann-Janette Rosga and Margaret Satterthwaite, for instance, identify the dangers associated with internationally agreed indicators 'artificially [closing] the gap between international law and domestic policy', whereby national political discourse and contestation about rights is simply bypassed through the manifestation of apparently neutral and objective indicators, agreed in supra-national fora, in domestic policy. 81 Indeed, the OHCHR's guide is explicit about this in its advocacy of the introduction of indicators into national human rights action plans, development strategies and budgeting. 82 National budgets being fixed in line with internationally agreed targets, whether one agrees with a given target or not, indisputably terminates or sublimates political conversation surrounding the use of public funds.

400
EJIL 27 (2016), 385-408 Yet in a sense, the closing-off of dialogue and conversation is a much more subtle and pervasive process, associated not with the outright and absolute introduction of human rights indicators wholesale into domestic policy but, rather, with the creation of a technique of global governance, to paraphrase Benedict Kingsbury and colleagues. It has been persuasively suggested that indicators are part of what Nikolas Rose refers to as a method of 'governing at a distance' 83 -a representation of how political power becomes reconfigured so as to provide for the autonomy of individual actors while, at the same time, governing their conduct through supposedly neutral measures -'the instrumentalization of a regulated autonomy'. 84 Governing at a distance, in other words, is governance that does not regulate actors directly but, rather, attempts to link their behaviour to political objectives and is particularly useful -or, indeed, crucial -where systems are decentralized. Indicators in the sphere of international human rights, it seems, are a mechanism for achieving this. They allow state actors to have autonomy within a largely decentralized system while, at the same time, aligning their behaviour towards certain political goals (that is, predetermined conceptions of human rights) through the use of apparently apolitical measures and standards. 85 Thus, the cloak of neutrality is of particular usefulness to the marketing of indicators. It is precisely because they depoliticize what would otherwise be highly contested choices that they are most useful. As Rosga and Sattherthwaite point out, indicators purport to offer technical answers to what would otherwise be extremely difficult questions, and this makes indicators presentable as a kind of neutral exercise that ought to be voluntarily accepted, rather than an assertion of authority and power. 86 This is useful in securing voluntary compliance on the part of what are commonly referred to as 'human rights stakeholders' and in persuading states to submit to the supervision of treaty bodies and the OHCHR on the basis that this supervision is merely technical and therefore non-threatening. However, by extension, this can only have the effect, and, indeed, is often designed to have the effect, of removing human rights from the sphere of domestic conversation and debate, which, in turn, distances human rights from the very stakeholders who have most at stake -ordinary citizens. 87 The language and mentality of auditing reinforces this denuding of discourse. For debate is not, of course, entirely ended by the adoption and creation of indicators. Rather, much remains to be discussed both in their creation and in the application of the broader conceptual framework. But this discussion, taking place as it does away from domestic politics in the realm of international experts and technicians, submerges political questions in technical arguments surrounding how and what to measure, how to collect data and so forth. 88 While political considerations are immanent in the process, they are left unexplored. Moreover, because quantitative measurement produces a sense of definitiveness and objective reality in the indicators it produces, it conceals the underlying socio-political claims the indicators may represent. 89 That is, in deciding what to include or exclude in a given indicator, the producer is making decisions, implicitly, about what is or is not desirable, and such a decision can only be a political, value-based process, especially in a field such as human rights. Depoliticizing what would ordinarily be politically contested, through the application of technical standards, may provide a sense of neutrality and objectivity, but it is a false sense. It removes socio-political values from the public realm and embeds them in the construction of indicators, which shifts the balance of power towards the experts engaged in that process. 90 The scope of influence of politicians and politics is correspondingly reduced, and the focus moves away from the ethical towards manner -from content to form. 91 Thus, technique replaces the difficult process of moral and political education that Oakeshott held so important: 'The morality of the Rationalist is the morality of the self-conscious pursuit of moral ideals, and the appropriate form of moral education is by precept, by the presentation and explanation of moral principles.' 92 This can be seen very clearly in the way in which indicators are intended to transform the administration of human rights law at the domestic level into, essentially, an exercise in discourse-free checking. Once the indicators have been agreed, human rights performance can apparently be assessed objectively, with quantitative improvements across the various metrics demonstrating that compliance is taking place and the human rights situation is improving. The questions of whether it actually is improving, and what this means, are sublimated beneath the auditing process, but, more importantly, so are questions of morality and the field of moral education. Rather than the slow, flexible evolution of the moral traditions and habits of a community through political conversation, there is instead simply the blunt application of the precept. Rather than a concept of a given right developing organically within the society in question, there is instead the displacement of localized rights and rights dialogue, 93 and the suffocation of local idioms. 94 Instead of injustice being framed around the individual, it is framed around the failure to perform against targets or the mere failure to produce auditable performance itself. Instead of being a discourse about establishing what is right and wrong -or just -in given cases or in a given policy, the matter of enshrining and protecting human rights instead becomes a task for management. 95

402
EJIL 27 (2016),  Ultimately, this also has the effect -and, as has been often seen, the stated intentionof diminishing the role of the subjective, experiential and practical in the protection of human rights and the assessment of its performance. It represents the reduction, as Oakeshott put it, of 'the tangle and variety of experience to a set of principles'. 96 The rationalist view of experience is a limiting one. It accepts only experience that has been converted into a formula, not the wider sense of experience cumulating within a society into traditions and habits that, though mysterious or opaque, contain its accumulated practical knowledge. This leads the rationalist to cut himself off from tradition and, in turn, become 'apt to attribute to mankind a necessary inexperience in all the critical moments of life'; 97 to paternalistically ignore local and contextual wisdom in the name of certain and uniform objective measurement -to seek to extract knowledge from its context and its history. 98 The power of indicators to reduce and submerge the complexities of contextual phenomena into standards and categories is a ready example of the capacity for rationalism to act in this way, as is the reduction of subjective, expert narrative into quasi-objective measures through coding into numerical representations or through the reduction of lived experience to survey data, which transforms the rich tapestry and complexity of communities and individuals into quantitative measures such as 'percentage of individuals who feel safe walking alone at night'. It is the imposition of the simplicity of principles onto the 'tangle and variety' of lived experience. It is to ignore how matters are currently 'attended to' and, instead, to approach matters as a social engineer. 99 Removing moral discourse from the sphere of human rights and driving it into 'what is measurable instead of what matters' then has the effect of depriving the human rights movement of its power as a mechanism for justice, and diminishing its potential by focusing on what is measurement friendly. 100 The unpredictable, ungovernable, chaotic nature of changing social norms, which may seem dormant for decades before undergoing dramatic shifts, is lost amid the need to demonstrate structure, process and outcome. Denuding the human rights movement of discourse by shifting the emphasis towards verification and auditing strips away its political power. The end result is the draining away of its spirit. Instead of Immanuel Kant's moral politician, the 'genuine republican' seeking to act in the right way based on the context of the society in which he lived, there is instead simply the development and application of professional technique. 101 Instead of narrative, the telling of stories 'that keep alive and strengthen the ideas of freedom, equality and universality', there is a managerial sensibility that sees its role as the verification of performance towards outcomes agreed among a caste of international experts, rather than justice. 102 96 Oakeshott, 'Rationalism in Politics', supra note 32, at 2. 97 Ibid., at 3. 98

The Transformation from Tool to Paradigm
Yet most advocates of the use of quantitative data in the measurement of human rights see a continued role for the subjective and experiential complemented by the use of indicators. The OHCHR's guide itself describes indicators as a 'tool to support' narrative or judicial assessments, and this perspective is fairly common. 103 As Nancy Thede describes it, the majority 'see the usefulness of a combination of analysis and quantified data, the latter being a useful complement to the former but never under any circumstances being able to stand on its own.' 104 And there are, indeed, alternative visions for the use of indicators that suggest how the technical knowledge provided by statistical measures can supplement the intuitive and tacit practical knowledge of the practitioner, rather than supplant it. An extensive example of such a process is provided by Christopher Stone in the form of 'locally usable near real time "active indicators"', used by the Jamaican Constabulary Force to mine data regarding raids and searches in order to establish best practices, or by the Attorney General of Lagos State to establish how many prisoners were on pre-trial detention in prisons and for how long in order to focus resources on the elimination of bottlenecks. 105 Here, quantitative data in the form of indicators is used in a decentralized fashion by local professionals and experts to notice patterns and find hidden discrepancies or misallocations of resources in order to improve their work. Stone contrasts this with the many 'carcasses of failed indicator projects' littering developed countries -the results of resistance to the exercising of external power on the part of local agencies ('those in positions of immediate and legitimate authority', as Stone puts it) and of constant shifting priorities forced by new indicators demanded by governmental ministers, donors or UN agencies. 106 Rather, Stone describes the use in Jamaica and Lagos of indicators that are designed specifically for use by an official with formal authority over those expected to produce the outcome being measured -for instance, the commissioner of the Jamaican Constabulary Force. 107 This design enables close, responsive, timely measurement for a clear purpose, improving the effectiveness and efficiency of police raids and, crucially, ensuring that data is examined at the level of operational responsibility by those in positions of legitimate authority, not national or international figures who are far away from the actual practice in question (both literally and figuratively) and who lack accountability and, hence, legitimacy in the eyes of those working 'on the ground'. Other examples from the Philippines and Colombia demonstrate how human rights indicators can be developed and used locally, even when they are created in response to international standards or principles, to provide information about local needs. 108 EJIL 27 (2016), 385-408 These, it seems, could be examples of Oakeshott's description of the combination of the technical and the practical made real: the tacit, intuitive, experiential nature of the practitioner supported and enhanced by the application of statistical tools in a bottom-up, contextual process carried out at the most decentralized level possible -and the approach certainly holds promise in the field of human rights. Given this possibility, then, it seems churlish, even disingenuous, to describe the move towards indicators as being in danger of crowding out less technical methods of monitoring human rights. Construed as simply being supplementary to, and supportive of, judgment-based, subjective assessments, where is the harm in an increased use of indicators, moderated by common sense?
The danger is that the rationalist tendency is always at risk of becoming dominant. Perhaps the greatest concern is simply that the desire for ever greater acquisition of technique will accelerate and that it will consequently come to subvert the experiential, practical knowledge, analysis and assessments of the locally embedded expert. Despite best intentions -in other words, despite human rights practitioners taking the view that quantitative measures and indicators are simply complementary to human judgment about how to best protect human rights in a given context -it may be that the rationalist tendency will nonetheless turn monstrous (not least because of financial considerations, which will always prioritize measureable 'impact'). Merry describes in compelling terms how, at the treaty body level, 'indicator culture' has come to be the default mode of monitoring human rights performance, 109 and in the UK, the Equality and Human Rights Commission's work has come to revolve heavily around the use of indicators in a 'measurement framework'. 110 At a recent workshop organized by the International Council on Human Rights Policy, the participants expressed the concern that 'evaluation methods and frames of impact assessment [are] exerting a disproportionate pressure on the narratives and practices of human rights work, distorting priorities, objectives and processes'. 111 Or, as Paul Gready puts it, '[f]or human rights, there is a real danger that its reference point may shift too radically from international human rights law to the requirements of evaluation and donors'. 112 Why should this be so? Oakeshott's final, and most pressing, concern was that rationalism would perpetuate itself -that the field of morality and moral education would become more and more a matter of the acquisition of technique -primarily because of the view the rationalist had towards practical knowledge. Since he saw it as being inferior to technical knowledge, he begins to neglect it, if not to attempt actively to extirpate it. And because of this, he 'dries up [his] mind' -he sees the only solutions to perceived problems as the application of yet more technique: ' All the Rationalist can do when left to himself is to replace one rationalist project in which he has failed by another in which he hopes to succeed'. 113 The political content of human rights -the goal of justice for the individual against the power of the State -becomes stripped away by the need for better measurement and better performance based on what can be measured. 114 The concern, then, is not that anybody advocates for the replacement of judgement-based assessment by the use of purely objective-seeming statistical measures, indicators and benchmarks. Rather, it is that the rationalist propensity sees the resolution of problems as purely being a matter for the application of technique, rather than for the more difficult, time-consuming and decentralized -uncontrolled -method of developing practical knowledge. Instead of the slow, steady, quiet evolution of moral principle and of the tacit, experiential knowledge of the expert, the default position becomes that quantitative measurement provides the cure for all ills. Once the culture of statistics and indicators enters the mainstream, it becomes difficult to reverse. Where practical knowledge is seen as 'nescience', the only solutions postulated will be technical in nature, and the only cure for rationalist failures will be more rationalism. And this describes with great accuracy the response to the problems associated with the indicator project, which are characterized by, if anything, an even greater tendency towards rationalism than the project itself.
Indeed, it seems that the temptation towards centralization and certitude in the production and use of indicators is almost inescapable once the indicator culture has taken hold. The rationalist tendency, unable or unwilling to leave too much space for the practical, for the experiential and for the tacit, always prefers the application of technique. This is most evidently the case for those who view the use of globally determined quantitative measures as being essential to the realization of human rights. 'While the criticisms and challenges to the creation of international indicators for ESCR are informative', as Welling rather revealingly puts it, 'they do not justify allowing the international community to remain uninformed about the deficiency of ESCR protection and access around the world.' 115 In other words, indicators may have their flaws, but the solution can only be the generation of more data that the 'international community' can utilize. The cure is only ever technical in character; it is the production of frameworks and systems by which techniques of indicator creation and usage can come to be implemented. The alternative is for the 'international community' to be uninformed.
Thus, the first step in the next stage of rights-based development is identified as 'the compilation of an internationally agreed and scientifically adequate list of core development indicators for civil and political rights, based on international standards'; 116 the solution to conceptual problems surrounding economic, social and cultural rights is presented as being the development of more detailed suites of indicators by the 406 EJIL 27 (2016), 385-408 CESCR so as to give 'concrete meaning' to those rights, 117 and the solution to discrepancies between agreed minimum standards for humanitarian relief due to differing local contexts is presented as '[describing] the gap between the standards and indicators listed in the handbook and the ones reached in actual practice … [and explaining] what needs to be changed'. 118 The OCHCHR's guide, which acknowledges the limitations of the use of indicators in its opening pages, argues that the solution is stronger involvement of human rights stakeholders in their production. 119 The problems associated with being half right are often resolved, that is, through the application of technique; through ever more rationalistic responses.
The tendency of rationalism to perpetuate itself is also apparent even where the entire indicator project is viewed with scepticism. Thede, for instance, while acknowledging the dangers of 'statistical mystique', the value-laden nature of indicators, the difficulty or impossibility of capturing values in numerical form and the prospect of a 'productivist' approach undermining human rights themselves, still comes to the conclusion that what is required is to 'establish what are the component aspects of each right, what the relationship of each component is to the overall respect of that right and, finally, what are the indicators that can generate meaningful information about one or the other of the components of that right' -in short, a 'theory to link concept to indicator'. 120 In other words, these issues are to be resolved with the application of reason to reach solutions, which is precisely the type of effort undertaken by the OHCHR to centralize and systematize a conceptual framework for determining and measuring the attributes of each and every human right.
Merry, meanwhile, in a careful and considered analysis of the problems associated with human rights indicators (noting that they are unrepresentative, dependent on the cooperation of states parties to the treaties, dependent on ranking to have any effect, decontextualized and ahistorical and lacking a sanction mechanism for failures of implementation) comes to the apparent conclusion that what is required is simply more 'powerful' indicators permitting ranking, punishment for lack of achievement and, above all, 'clear theory embedded in the indicator' -in other words, more simplification, more clarity and more streamlining. 121 The problems associated with the application of technique are cured by the application of yet more of it.
And despite an extensive and considered critique of the development of the indicator project, Rosga and Satterthwaite similarly propose a solution that boils down to the production of a set of outcome indicators produced by the treaty bodies, the OHCHR and NGOs, to be applied universally. Here, the approach adopted by Fukuda-Parr and colleagues, that is, 'ranking countries by measuring the relationship between the extent to which a population enjoys fundamental economic and social rights and the resource capacity of the state to fulfil ESR obligations' is cited as promising, though how this approach avoids the pitfalls identified by the authors is not made clear. 122 Political conversation at the domestic level is to continue, but only if it is framed around the universally applicable Donabedian model, with the power to determine structural and process indicators devolved to the national level -outcome indicators having been centrally mandated. In addition, indicators must be created to measure political participation so as to avoid the bypassing of political discourse in domestic politics. 123 The solution to the pitfalls of governmentality, it turns out, is more of it. What is required is the utilization of indicators to permit global governance to interface with the governed more directly. 124 The rationalist propensity's tendency to propose rationalist solutions is well illustrated by the fact that even Stone is still tempted, ultimately, to suggest that what is required is an 'iterative method' for the construction of global indicators. That is, since he is unwilling to entirely leave the production and use of indicators to practitioners or experts, whether liaising across borders or otherwise, Stone eventually comes to the proposition that there ought to be a 'system or framework of indicators that can be used across countries, in very different national contexts'. 125 While this ought to be produced from the 'bottom up', the character of the rationalist still lurks in the background in that the end result seems to remain a globally produced and universally applicable system. The rationalist propensity is unsatisfied with any approach that does not, ultimately, lead to the uniform and certain application of principle, however this is produced. While there is promise, then, in the notion of indicators produced and consumed by those in positions of legitimate authority, as Stone suggests, there must come with it a wariness of rationalist tendencies towards uniformity and certainty in different guises.

Conclusion
Describing the indicator project, and the indicator culture that surrounds it, as 'rationalist' gives us a fruitful method for understanding not only its flaws but also its growth. What, then, is to be proposed? One obvious critique of Oakeshott's position is that, as elsewhere in his work, he seems to present a bleak scenario in which matters have progressed too quickly and too far in the wrong direction to ever be reversed. 126 He appears to present us with a vision of a rationalist tendency that grows ever more pronounced until there is space for nothing else. This article considers that vision to be compelling but, at the same time, recognizes that there is considerably more nuance