The Governmental Topologies of Database Devices

Topological Analytics

I first conceive of these databases as ‘devices’ as they are government schemes devised to both calculate and intervene in the performance of individuals and populations, and consist of an ensemble or system of relations. That is, they are oriented to a governmental purpose and, as such, the terminology is most closely aligned with what Foucault defined as dispositif:

a heterogeneous ensemble consisting of discourses, institutions, architectural forms, regulatory decisions, laws, administrative measures, scientific statements, philosophical, moral and philanthropic propositions – in short, the said as much as the unsaid. Such are the elements of the apparatus. The apparatus itself is the system of relations that can be established between these elements. (Foucault, 1980a: 194)

In this light, database devices consist not only of inscriptions but strategic orderings and relations between many governing actors and elements. Actor-network theory (ANT) is helpful in specifying and advancing this understanding of relations by attending to the role of ‘immutable mobiles’ in enabling such arrangements to ‘hold’ in some relational or functional manner between sites, and to form more or less stable networks (Law and Singleton, 2005). But what often happens is that devices (their elements or relations) tend to change and the network ‘fails’. A solution that is often advanced is to explore how a network struggles to maintain the stability of its devices. But this means that invariance becomes only a principle because in practice devices can be mutable and multiple (Mol and Law, 1994). ⁶ Topological analytics provide a way of thinking of devices that accounts for such instabilities of devices and the mutability of their elements, relations and boundaries.

Rather than a given theory, social scientists generally identify a few key properties of the topological from mathematics and have then taken these up to study social phenomena. One property is a non-Euclidean and non-topographical conception of space, which Celia Lury (2009) has described as follows:

Topology in mathematics does not start with a space but starts with a problem (an equation) then explores the space in which it has a solution. So while geometry had classically been understood as the perception and organisation of a static, homogenous space that might be projected reliably into the future, topology offers a way of thinking about processes of actualization in n-dimensional spaces in terms of probabilities, not certain futures. To put this another way, space, problem and solution are co-constituted in topological mathematics. If this is indeed so, then it suggests that what is involved in thinking topologically across disciplines is neither induction nor deduction but abduction. And abduction does not involve either metaphors or tools – but speculative reason or reasonable guesses.

Thus, rather than static and predictable, a topological space (or object or subject) is mutable, its boundaries changeable and modulating in unpredictable ways giving rise to probabilistic rather than predictive reasoning. Law and Singleton (2005) state this differently: that objects have no clear boundaries and are made up of heterogeneous mixtures of variable elements with unstable relations that can be discontinuous, unpredictable and generate varying patterns of absence and presence. Yet, as Law (2002) notes, while some properties are changing others are continuous and retained. The challenge of a topological analytics is to identify the possibilities and properties of different forms of continuous transformation, that is, ways of deforming objects while securing their continuity.

Drawing on these understandings, I conceive of topological analytics as involving a series of interrelated presuppositions: the co-constitution of a space, problem and solution resulting in varying rather than fixed boundaries and a logic of abduction or reasonable guesses; and invariance alongside varying and heterogeneous elements and relations that can be folded and reordered and generate shifting patterns of present and absent elements. These properties form the analytic starting point for conceptualizing the ontology of the subject and governmental logic of database devices. To start, I specify the content of government databases as being made up of transactional metrics, and then I take up these topological analytics in relation to a number of social policy initiatives and database devices.

Population Metrics: The Stable and the Transforming

Government administrative databases constitute a particular kind of measurement and data on subjects. They consist of past, current and ongoing measurements of conduct recorded through established means of data collection (e.g. tax, national insurance or school records). The databases track the transactions of individuals in relation to government services: their registration of life events, income earned and taxes paid, licences obtained, schools attended, cars purchased, borders crossed, visas acquired, benefits received, visits made to hospitals and so on. ⁷ They also contain records of ‘encounters’ with governments and log changes in individual lives. The data is generated as a by-product of everyday administrative transactions with government and is deemed ‘better information’ in part because it is considered dynamic measurements of what people ‘actually’ do and need, which change over time and space and thus require constant monitoring and evaluation. ⁸

To be sure, transactions have long been the basis of government administrative systems and the mainstays of government record keeping, monitoring and evaluation, and, in social policy, conduct and behaviour have been key indicators and registers of identity. What is new is the digitization of these records, the possibility of real-time tracking of transactions and the potential to join up transactional data distributed across government sites and functions. In other words, digitized data are transforming government knowledge practices.

Administrative databases and the identification of correlations and patterns in transactions across governing sites constitute particular kinds of operation and measurement that I call population metrics – measures that identify and evaluate the performance of individuals and populations (Ruppert, 2010). Population metrics work in much the same way as metrics operate in business, government and academia, to evaluate and compare the performance and progress of people, groups and things. For example, in education, league tables and scores evaluate schools, in universities, bibliometrics measure academic performance and in health care, standards such as wait times evaluate service delivery. The same logic applies to how transactions are used in government to identify individuals and populations and evaluate their performance.

But before transactions can be combined, reassembled and correlated, ‘identity management’ is required to confirm, verify or authenticate identities, prove eligibility and entitlement to public services and benefits, trace and track transactions and movements, and join up databases across sites and functions. In addition to assigning a unique identifier or code, identity management involves recording what are considered relatively stable biographical identifiers and locators that can include name, date and place of birth, gender, and address and biometrics. Identity management thus involves the stabilization of a set of biographical and/or biometric classifications or ‘single source of truth’ about individuals that is kept up to date (Varney, 2006: 38).

For the UK government, identity management is achieved through the use of unique identifying numbers and codes. There are numerous such identifiers in circulation: national insurance, pupil, health, and driver licence numbers. ⁹ Once identities are managed in this way, then individuals and populations can be known in relation to their multiple transactions and movements.

Identifying: Multiplicities and Singularities

Identity management constitutes an understanding of individuals as being made up of a set of unchangeable and stabilized identifications. Once identity management is in place, then it is possible to track, trace and infer knowledge about individuals and populations through the linking of multiple transactions, which become the dynamic and relevant metrics to be measured, monitored and analysed. This understanding of individuals was most comprehensively laid out in the UK Labour government’s ‘Transformational Government’ strategy, which sought to use new technologies to transform the ‘business of government’ by joining up and sharing data and services rather than duplicating them. Multiple sources of information are deemed necessary to provide complete pictures of customer needs and behaviours, which are multiple and complex (Varney, 2006). People do not fit into the usual group categories (the elderly, the student) but rather – from the governing point of view – have changing and multiple needs that require ‘responsive services’ (Cabinet Office, 2006b: 8).

It is this conception that informed many specific UK social policy initiatives. ¹⁰ The Social Exclusion programme was based on the understanding that relative disadvantage is caused by linked, multiple health and social problems and risk factors that sometimes ‘interact’ to multiply their overall effects (Cabinet Office, 2006a). A key solution promoted was the joining up of databases to identify individuals at risk of social exclusion, such as those who are ‘poverty plus’ or have ‘chaotic lives and multiple needs’ (2006a: 76). Database devices that materialize this conception of the joined-up subject have been most significantly advanced and implemented in two key areas: child welfare and youth justice. I will thus focus on these examples in the remainder of the article. ¹¹

The Integrated Children’s System (ICS) – an e-social care record system and database for managing information on children who are or may be in need – is designed to collect and record a great amount of detailed information from different practitioners (social workers, head teachers, counsellors, police, health care workers and so on) and to predict children who may be at risk of abuse or potentially in need (Department for Children, Schools and Families, 2007). The data are recordings of transactions and interactions between practitioners and children and their parents, including contact information, assessments, service planning and service review. The logic is that, by joining up data, children can be identified in relation to a combination of needs, an understanding that no one practitioner could ever possibly attain because the subject’s relations and interactions are distributed among several service points. It is only by connecting and integrating the data that a child in need or potentially at risk of abuse can be made visible. Each child’s combination of singular transactions with government agencies, when categorized (e.g. failing grades, non-attendance at school, criminal record), produces a metric or measurement of needs along a continuum from no additional needs, additional needs to complex needs.

The Youth Justice Board Management Information System (MIS) involves a similar conception of young people. Each local authority in England and Wales has a multi-agency Youth Offending Team (YOT), with representation from the police, the probation service, social services, the health service and education. ¹² YOTs enter cross-agency individual-level data into the system and the data is then used to assess children and young people (10–17 years of age, what I will refer to as ‘youth’) and their likelihood of (re)offending. The MIS assembles biographical and transactional data (evaluations, assessments, interventions, judgments, sentencing) of young offenders compiled across distributed government sites: schools/colleges, police, general practitioners, health service providers, social services, housing, voluntary organizations, courts and so on. The data includes case-level information compiled by myriad practitioners (police, housing officers, social workers): offences details, court data (charges, sentencing), intervention records (e.g. assessments, plans, dates of contacts, outcomes), assessments of assets at various stages of contact (12 dynamic factors that are scored, such as living arrangements, education, lifestyle, substance use, emotional health), indicators of vulnerability and risk (based on a series of questions), restorative justice interventions, parenting interventions (meetings) and mental health and substance abuse interventions (referrals, assessments, treatments). ¹³ Collectively, these constitute a series of metrics that measure the performance of youth in relation to myriad government services.

The two programmes and their related database devices are based on the logic that the subject is made up of unique combinations of distributed transactional metrics that reveal who they are and their capacities, problems and needs. An individual is not simply a child or youth, but rather a combination of needs and services. The population to which they belong is thus based on associations not so much with others but patterns in their transactions. Existing disadvantage, for example, is detected in relation to combinations of transactions for benefits, employment support, health care, mental health services, alcohol services, housing and supported housing, homelessness services, policing, prisons, courts and drug services. Children and youth are discovered through combinations of multiple transactions that vary and change. In other words, new and emergent subjects can be discovered and made up by these technologies: the youth at risk of (re)offending or the child with complex needs. While similar kinds of subjects may have been previously identified, database devices introduce a change in criteria, detection and reporting that capture new people as part of these populations. Furthermore, they intensify individualization by assembling ever more and diverse metrics to specify subjects while identity management secures their continuity within their transforming and changing constitution.

Thus, rather than abstractions or disembodiments, these constitute particular materializations of both subjects and populations. The transactional metrics of subjects and their performance in relation to practitioners constitute a kind of ‘surface’ rather than determinist or causal understandings (Savage, 2009). But it is through the assembling of data distributed across distant government sites that patterns and associations in the performances of subjects are detected. Government sites and their contained data are thus ‘folded’ and brought closer together by and into the database device. Additionally, rather than transporting immutable mobiles to different sites to stabilize forms and relations and exert control from a distance (as conceived by ANT), it is an assembling device that gathers diverse and distributed elements constituted by heterogeneous arrangements and practices.

As such, the boundaries of government databases become permeable and what is assembled – included/excluded, present/absent, inside/outside – of the database device can vary and change. What ‘makes up’ the subject and the population, or what could be called the boundaries of the social, is thus mutable as varying metrics can be joined up, arrayed and correlated across sites and databases. Such arrays do not assemble all data but do so selectively and thereby create unstable patterns of absence and presence. Yet the biographical metrics of identity management provide a ‘core of stability’ (Law and Singleton, 2005) that enables governing sites to be brought close together. What then is mutable is the transactional metrics and relations between performances that make up the subject, and for social policy these are key matters of concern. ¹⁴

In this regard, database devices are materializations of what Latour (1998) has called a traceable social that is rendered visible, not by extracting it from something else but by making it observable. ICTs are usually depicted as communicating information stripped of the encumbrances of social relations and location (Strathern, 2000). Rather, ICTs do not abstract and detach from the social only then to be put back into it but are part and parcel of the very relations that get materialized in data. Rather than occupying a ‘space of flows’ or a virtual informationalized world, data is itself a materiality that can be assembled in multiple ways by folding in heterogeneous data from distributed sites.

Such traceability does not reveal an overarching society but a multiplicity of links and orders made up of aggregations of individuals. In relation to populations such materializations are of course not new. Governments and social scientists have long produced materializations, from censuses and maps to other inscription devices. As such, the point is not that materializations are becoming more abstract but that social orders are being materialized in new ways and in specifically more traceable ways that are enabling the making and identification of new connections and the capacity to see populations being made and remade, to see populations as being composed and recomposed.

But so too is the individual subject being materialized in intensified ways, as a monad, a being composed of a ‘vast crowd of elements’ (Latour, 2010: 10) that only joined-up databases and their visualizations can identify. ¹⁵ Both the individual and the population are beings beyond human perception and thus database devices are required to mediate and make them visible. What are those elements? They are the interactions and transactions of a child with government agencies and her/his identification as a child in need – that is, each child is an interiorization of a whole set of relations, interactions, assessments and evaluations, or what I have called metrics. The child does not occupy one subject position nor multiple subject positions but, in relation to government, is a multiplicity. And the population is an aggregation of individuals who are themselves also aggregations of multiple elements. What is assembled then is not the product of any one relation or performance but numerous ones such that the enactment of both the subject and population is more precarious, indeterminate and unpredictable.

This conception of the individual as a monad is suggested in the Transformational Government agenda (Cabinet Office, 2006b). It advanced the understanding that people rarely fall neatly into categories and associate themselves with different groups at different times depending on their particular needs, and so interventions must be responsive to their unique needs (Cabinet Office, 2005). Each person is conceived of as a singularity; not a member of a group or having a generalized group identity, but made up of a singular composition of performances that need to be guided and shaped. Database devices both materialize and intensify this composition while at the same time enabling it to mutate and change.

Governing: Singularities

In sum, patterns among transactional metrics are measures of the relative performance of children or young offenders who are not placed into categories of normality or abnormality. Categories are the disciplinary techniques that Foucault first described, which generally operate through classification schemes and statistics that categorize and organize individuals into populations – the poor, the criminal or the delinquent – and govern through what Deleuze (1992) also described as spaces of enclosure such as the school and prison. However, as Foucault (2003, 2007) argued in his later lectures at the Collège de France, such technologies of discipline that concentrate and enclose give way to a positive power and technologies that expand, where ‘[n]ew elements are constantly being integrated: production, psychology, behavior’, ‘allowing the development of ever-wider circuits’ which ‘lets things happen’ (Foucault, 2007: 45). ¹⁶ It is an inventive power that works by inclusion and enables transformation and innovation through the interplay of coexisting government interventions (education, welfare, police and so on) that get materialized by the database device and ‘whose effect will be greater than the sum of its component parts’ (Deleuze, 1992: 3). That is, rather than a simple operation of addition, the database device is generative of new and emergent subjects and populations.

Importantly, it is through ‘fine-grained’ individualization and what Foucault described as ‘a series of fine and constantly observed differences’ (2003: 46) that differential normalizing is achieved. This is the ontology of the subject that the database devices materialize, as a monad made up of complex, unique, dynamic and always varying metrics. While children can still be generalized into patterns of likeness or similarity, immense detailed data on them is maintained and linked to a pattern. In this regard there is multiplication and complexification rather than simply the generalization of the individual into a category. Rather than a series of categories, children are located along a continuum of need where various services and interventions can be said to ‘co-exist in one and the same modulation’ (Deleuze, 1992: 5). Amoore and de Goede (2008) develop a similar argument in relation to security practices where transactional data is used to classify people not in relation to categories but according to degrees of differential risk. At the same time, such devices seek to ‘hold’ the individual in relation to the whole population by maintaining his/her specificities and variations, and making these traceable and visible. The population is the aggregation of these specificities and variations, such that each child occupies a singular point within the pattern of needs of the 370,000 children in England and Wales (statistic cited in Shaw et al., 2009).

These devices thus work vitalistically – they aggregate individuals to assess the relative health of a population (the number of children with different degrees of need or the number of youth at risk of (re)offending). This is of course the general problematic of governing, which is to know the nature and then govern and regulate the forces of the whole, that is, the population, the referent object of biopolitics (Foucault, 1997). It requires specific totalizing procedures – that is, techniques that can constitute an entity out of various individual parts – all the rates, profiles, patterns and probabilities necessary to manage, regulate and maximize the potential of a changing population (Dillon and Lubo-Guerrero, 2008). The social policy programmes mentioned above and the database devices in question seek to do so and to maximize populations: to identify and target a population of people who are socially excluded or at risk of being so; and to protect children by sorting them into categories of need and intervening in their development. Such identifications are then the basis of governing which seeks to prevent an increase in the number of certain kinds of people (socially excluded, abused child, teenage mother, youth offender). ¹⁷ That is, once identified as a probability – children at risk of abuse or youth at risk of (re)offending – then individualizing interventions can be defined. In this regard, susceptibility is not identified following a logic of induction or deduction but instead abduction: through patterns in joined-up data probabilities, reasonable speculations and precautionary principles lead to the identification of possible kinds of people.

In this way biopower and governing are integrated: the former track, regularize and manage populations and the latter guide and shape individual bodies (Foucault, 1980b). For while identification is based on a person’s metrics, interventions consist of tailored and individualized rather than collectivized strategies. The individual may be generalized to become part of a population (totalizing) but at the same time her/his singular identification or ‘fine-grained’ individuality is maintained, thus opening her/him up to singularized governing interventions (individualizing).

This is the logic of ‘modernising of public services’, which involves defining personalized packages of public services (Office for National Statistics, 2005). People are conceived as passive recipients of packages of services formulated from distributed data about them and targeted to meet their needs but not seen by them. ¹⁸ It is multiple transactions and their performances in relation to government services that reveal their problems and needs. Metrics are not based on a causal model but on patterns, regularities and a surface of interfaces and connections between measurements of what people do.

But targeted government interventions further generate varying and unstable relations. Metrics are a ‘species of nominalism’ (Hacking, 2007: 294) whereby being classified and named a ‘child with complex needs’ is only part of the dynamic. In addition to the experts and technologies of identification and their diagnostic practices, bureaucratizing practices then intervene to guide and shape and sometimes ‘correct’ people. Ian Hacking, in his many writings, has argued that both processes are involved in ‘making up people’: a new metric can bring into being a new conception and experience of a way to be a person. Furthermore, there is a ‘looping effect’, a process by which a metric may interact with the people through governing interventions that reinforce the ‘identity’ of a person so discovered (Hacking, 2007). ¹⁹ Because of these two processes, people are moving targets: metrics interact with people and change them and since they are changed they are not quite the same kind of people as before. Thus as people change metrics also change and become modulating such that populations are in a constant process of differentiating. Database devices thus also evolve alongside the performances of individuals and are not separate from them such that the technique and people are co-constituting. Hence, rather than a fixed shape and pattern of social structures and differentiations, populations are modulating differentiations. This challenges the usual way that population is understood, as consisting of people with different traits and behaviours that can be identified, categorized and governed. But this also means that governing interventions must also modulate and change in relation to ‘moving’ people. Not only must interventions be personalized and individualized, they must also change in relation to the changes that they have in part constituted.

Governmental Topologies

The foregoing is a conception that does not simplify but constitutes what Deleuze called a ‘complicating machine’ (Rajchman, 2000). But it is not complicating because of the availability of large volumes of digital data, the computational power of computer technologies and the myriad socio-technical relations that make them up. Rather, it is complicating because database devices materialize and intensify a conception of subjects and populations as always coming into existence through sets of modulating transactional metrics and relations where change is immanent in conduct. It is transactional metrics that differentiate and are productive of who people are and it is through their performance, or doing, that individuals and populations are registered and identified. The usual categories of sociological and social policy interest such as ‘group identifications’ or ‘backgrounds’ (e.g. ethnicity, religion, class, gender) and causal and depth models give way to ‘surface’ phenomena such as descriptions of patterns, links and regularities in conduct (Savage, 2009). The former are relatively fixed (date of birth, place of birth), or change slowly and predictably (age), or can be captured in subjective identifications (ethnicity) and constitute the relatively immutable elements of identity management. However, transactions change, and so metrics modulate as they track, measure and evaluate performance. The transactions of individuals dynamically constitute population, and it is relations between rather than detailed descriptions of transactions that matter. The object of interest is thus not the substantive elements of culture but their links and transformations, not some essential properties but instead relations (Lash and Lury, 2007). So while capturing identity preoccupies techniques such as censuses, joined-up administrative databases locate differences in relation to multiple registers of conduct that define both who people are and who they are possibly becoming. In this regard, population metrics are akin to what Rogers (2009) has called ‘post-demographic’.

In the introduction I specified a set of properties of topological analytics for understanding the database devices described above and the subjects and logics that they enact. With these analytics I argued that they constitute a social space by assembling and folding in dispersed data to generate heterogeneous and changeable mixtures of varying metrics and relations. At the same time, they enact individuals as sets of stable properties or identifiers such that subjects can be traced and tracked and understood as always in the process of changing and becoming. Transactions are the transforming and informing of who people are and their joining up constitutes metrics that evaluate and measure the performance of individuals and populations. But metrics and governing interventions are also dynamic and redefined and reformulated and deformed as a consequence of the looping effect between names and the named.

But rather than being ‘new’ it is an ontology of the subject and governmental logic that perhaps only now can be materialized and intensified through and with a database device. For Latour (2010), it is the ontology of individuals and the social that was advanced by Gabriele Tarde in the 19th century. While Durkheim sought to understand the individual in relation to the whole of a society and structuring social laws, Tarde sought to follow individual monads and their formation into aggregates and multiple social orderings. Latour argues that it was in part a lack of information and the incapacity to grasp the multiple and complex qualities of individuals and their aggregations that resulted in the dominance of the Durkheimian understanding rather than the Tardean conception of the social. Now, with the proliferation of digital data and devices, he argues that Tarde’s conception can be realized. I am not interested in the Durkheim versus Tarde debate, nor in recovering Tarde for sociology, but rather how this understanding is useful for interpreting the relation between devices and governing logics. Database devices make it possible to materialize a conception of population as a space of relations consisting of multiple aggregates of individuals with fixed metrics (biographies) along with complex and always-varying ones (transactions, conduct). This is not a conception determined but rather intensified and materialized by information technologies. Similarly, it is a materialization and intensification of what Foucault called the ‘fine grain of individuality’ and the singularity of the case (2003: xxii) and of a government logic elaborated by Deleuze (1992) as modulating control.

To conclude, by deploying topological analytics I have argued that database devices enact and advance a particular conception of subjects and governmental logic. Other analytics are certainly possible. For example, in relation to youth justice database devices structuralist analytics advance the argument that risk assessments obscure and aggravate existing social inequalities; actuarial analytics identify the problems of individual interventions based on inferences and population-wide probabilities; managerialist analytics establish the shortcomings and failures of risk criteria and scales; and political analytics identify discriminatory and human rights consequences of risk assessments based on what youth might do. ²⁰ I have also noted different social theory analytics such as Foucault’s dispositif or Latour’s actor-networks. But rather than offer topological analytics as an alternative perspective, I adopted an ontological stance to argue that different reals are enacted by different analytics. It is a stance that understands phenomena as already and always multiple and complex, and does not deny other possibilities but positions these as questions of ‘ontological politics’ (Mol, 1999). All analytics order and enact the world in ways that make some elements and relations present while absenting others (Law and Singleton, 2005).

What elements and relations do topological analytics make present? Principally, they emphasize variable, unstable and modulating relations, uncertainty, permeable boundaries and abduction as immanent rather than exceptional qualities of database devices. While there have been numerous critiques of neoliberal strategies, little attention has been paid to how databases and new analytics are advancing an ontology of subjects and populations as sets of unstable, transforming and generative transactional relations, and likenesses and identifiers of who they are and ‘of a potential future person yet to come’ (Amoore, 2009: 18). ²¹ Instead, social science analytics remain fixed on categories of identity and techniques of better knowing and governing the subject. However, not only government but also commercial, social and political practices involve the enacting of multiple forms of association and identification that are more variable, unstable and modulating than ‘older’ forms of identity. This is what topological analytics capture. Government database devices intensify these qualities through what Foucault described as dividing practices that extend through ‘ever-wider circuits’ (2003: 45) and ‘series of fine and constantly observed differences’ (2003: 46). Such differential practices are materialized and intensified by database devices, which are constituted by and generative of uncertainty and instability in both how the subject is known and governed.

To conclude, if managerialist analytics attend to how governing practices are deterministic and aim to stabilize forms and relations, then topological analytics start with the assumption that ‘the world is messy’ and that ‘we cannot know it by insisting that it is clear’ (Law and Singleton, 2005: 350). Or, as Mol et al. have stated in relation to care practices, ‘we do not bracket failure and fragility, but face up to them’ and learn to live ‘with the erratic’ (2010: 10, emphasis in original). By adopting these presuppositions I am not claiming that topology offers a better analytic (and indeed I have emphasized that analytics are not perspectives on but enactments of objects and subjects). Instead, it makes qualities such as fragility and uncertainty present and thereby opens up a different line of inquiry and set of questions about the governing consequences of database devices. Here I will suggest a few. If complexity, uncertainty and instability are understood as intrinsic rather than indications of system failure, then what does this mean for practitioners who must take decisions? Do these qualities demand more rather than less human judgement and intervention? If the identity of people is mutable then does this open up or close down the capacity of subjects to intervene in their identification? Are they rendered more passive or does uncertainty afford them opportunities to challenge or confound identification? ²² In other words, topological analytics turns our attention to questions of how practitioners and subjects work with rather than seek to tame complexity and instability and the possibility that these elements are not exceptions but givens and part and parcel of a governmental logic and ontology of subjects.