, Volume 3, Issue 1, pp 1–16

Nanotechnology, Contingency and Finitude


    • ESRC Centre for Business Relationships, Accountability, Sustainability and SocietyCardiff University
Original Paper

DOI: 10.1007/s11569-009-0057-z

Cite this article as:
Groves, C. Nanoethics (2009) 3: 1. doi:10.1007/s11569-009-0057-z


It is argued that the social significance of nanotechnologies should be understood in terms of the politics and ethics of uncertainty. This means that the uncertainties surrounding the present and future development of nanotechnologies should not be interpreted, first and foremost, in terms of concepts of risk. It is argued that risk, as a way of managing uncertain futures, has a particular historical genealogy, and as such implies a specific politics and ethics. It is proposed, instead, that the concepts of contingency and of finitude must be central to any understanding of the ethical significance of nanotechnologies, as these concepts can be used to understand the basis of recent work in science and technology studies, and the sociology of knowledge more widely, which details the multi-dimensional social nature of technological uncertainty.




That the future is uncertain is a precondition of both political and ethical action. In recent years, however, the analysis of a ‘politics of uncertainty’ has emerged as a topic in its own right, based on the insight that the management of uncertainty is a ‘quality of society’, as each society ‘evolves its own forms of understanding the world and regulating human behaviour, and distributes both the power to control relationships and freedom of action unevenly between its members’ ([56]: 1).

These forms of understanding express different interpretations of the meaning of uncertainty, which depend in turn on the extent to which specific actions are thought to determine future outcomes ([56]: 16–17). How the perceived need to incorporate techniques for conceptualising and managing uncertainty into social practices has influenced the shape of forms of governance has become a key theme of contemporary social science research, with the role of technology both in the management and the production of uncertainty generating much interest. For example, industrialisation’s alteration of the relationship between human society and nature has been interpreted as a cause of increased anxiety over how technologies and globalised social and economic structures create and distribute uncontainable risks [11, 13], which combines with other social-structural elements to create a generalised condition of uncertainty [10]. Among these elements are patterns of governance which, in conjunction with new technologies, have created a new individualisation of uncertainty, placing the burden of managing it elsewhere than upon the state ([72]: 236–7, 246–7). A key theme of some recent sociology is how uncertainty, in various social contexts, increasingly gets defined as risk. Defining uncertainty as risk implies the creation of a new object of knowledge, a bounded, measurable sample of homogeneous events ([14]: 578–579), thus constructing the future as an object which is calculable within certain parameters ([30, 73]: 7).

The ethics and politics of uncertainty might there be said to concern on the one hand how certain meanings of uncertainty such as risk, together with their normative implications, are privileged at the expense of others [87], and on the other, how the hegemony of these meanings creates and distributes specific harms and benefits. With respect to the development of radical technologies, the role of risk discourse in organising power relations between governance/business elites, experts and the public, and in shaping the social priorities which underlie processes of technological innovation has been much discussed, particularly in relation to genetic engineering and GMOs e.g. [46, 57]. Nanoscale science and technologies (hereafter NSST), due to the uncertainties surrounding both their possible future evolution and the potential novel properties of nanomaterials themselves, have been widely framed as a test case for new modes of governance which incorporate novel forms of public participation in decision making, together with precautionary risk management approaches [47, 78]. It might be argued, therefore, that the development of NSST governance represents an opportunity to consider the ethical and political significance of uncertainty directly in relation to a novel technology in an early stage of development. I present here a contribution to such an analysis, based on an investigation of how uncertainty is constructed and framed as an object of knowledge within different dominant narratives about the future of NSST. The goal of this investigation is to explore how narratives about technological uncertainty constrain the options for collective action we assume are available to us in the present. In “Techniques, Technologies and Being Uncertain”, I examine how the use of risk and future narratives of expectation in interpreting uncertainty are buttressed by assumptions about the centrality of technique and progress to technological development. In “Mapping Contingency”, I construct a typology of uncertainties surrounding technological innovation and locate within this the interpretations of uncertainty discussed in “Techniques, Technologies and Being Uncertain”. Finally, I indicate how this typology places in question the overwhelming focus on risk which is a key element of debates over the potential impacts of NSST.

Techniques, Technologies and Being Uncertain

How social uncertainties are interpreted, filtered and managed by discourses which articulate the promises and potential hazards of new technologies are important features of technological innovation. Employing technical concepts of risk, or more metaphorical uses of ‘risk’ which draw upon them, is one particularly potent way of normatively defining what uncertainties concerning the possible impacts of new technologies are seen as important and worthy of investigation ([46]: 294–96). Another way of framing technological uncertainty is to construct a range of possible and plausible futures through narratives of expectation [15] that are designed to showcase the socially transformative capacity of the technology. In this section, I review how each of these different practices buttresses the other, and how this mutual support is provided through their common basis in assumptions about the relationship between present action and future outcomes. These assumptions will be shown to be structured by more basic metaphors which depict action as reducible to technique and represent the creation of futures through the use of scientifically-informed techniques as oriented necessarily towards progress. These metaphors already imply a particular interpretation of uncertainty as a condition which shrouds facts that are not currently known, but which in principle can become known. As we shall see in the next section, this is necessarily a partial definition of uncertainty. In fact, relying on narratives of scientific or technical change to explain why social change might take particular directions places the process of social change outside political debate ([20]: 200–201). Providing a normatively-charged definition of uncertainty as risk plays a similar role, by making the interpretation of uncertainty into, primarily, a game for experts.

In the case of NSST, as with other technologies such as xenotransplantation and gene therapy, narratives of expectation have played an important role in mobilising funding streams for research and commercialization. In the context of an emerging focus within science and technology studies (STS) on the role of specific forms of future-orientation in constituting practices and meaning, the nature of the relationship between NSST fact and fiction has been remarked upon by several commentators [53, 68]. The construction of possible NSST futures has played a key role in securing funding for programmes such as the National Nanotechnology Initiative in the USA, inspired by the ‘Feynman vision’ ([21]: 26) of nanoscale engineering. The dynamics of innovation into which such narratives feed generate areas of interest which over time either consolidate by opening up new fruitful areas of research and further funding opportunities, or decay through the failure to fulfil initial promises.

A key element of such NSST narratives is how they develop an implicit understanding of technological (and social) progress as the mastery of techniques of precise control, and how they deploy this understanding in order to construct visions of desirable futures, heralding a future of molecular engineering as the apogee of NSST development.

What could we humans do if we could assemble the basic ingredients of the material world with even a glint of nature’s virtuosity? What if we could build things the way nature does—atom by atom and molecule by molecule? […] Whenever scientists and engineers push their understanding and control over matter to finer scales, as they now are doing on the nanoscale, they invariably discover qualitatively new phenomena and invent qualitatively new technologies. “Nanotechnology is the builder’s final frontier,” remarks Nobel laureate Richard Smalley, Rice University. ([5], 1)

This influential early vision of future nanotech continues to inspire research which follows both self-assembly [63] and single molecule cut-and-paste approaches e.g. [65] to bottom up manufacture of nanostructures for various applications. Imagery of “nanofactories” and assembly lines gives shape to the impending future which may be built on these discoveries (e.g. [17], 978, 982).

On a more theoretical level, the concept of technique is a way of understanding action as the use of scientific knowledge of natural causal processes for intervening within those same processes. Information about the results of these attempts enters a feedback loop which leads to further refinement of the applied knowledge (and which may, by revealing more about natural processes, modify ‘pure’ theoretical knowledge). The aim is to fine-tune technical interventions so that they produce desirable and predictable results. This understanding of progress as the ‘fine-tuning’ of techniques provides support for an understanding of uncertainty as, primarily, risk. To perfect a technique for producing a given structure or structures would be to perfect the varieties of control which are needed in order to make structures that will endure in time for as long as possible without significant decay or destabilisation ([45]: 58). If a set of techniques were incrementally perfectible, then this would imply that the uncertainties surrounding their employment can be progressively removed or at least made increasingly quantifiable and manageable. The idea of technical progress could therefore be viewed as the concept of progressively increasing control over uncertainties.

This vision of progress can be traced at work within various manifestations of a predominant narrative about the future of NSST, one which has featured for example in the creation and ongoing development of the US National Nanotechnology Initiative, but has also been influential much more widely. In fact, the narrative arguably forms the basis for most depictions of NSST futures in both scientific and popular media. It sets out several stages of incremental technical progress, beginning with the current stage of development of NSST, where nanoscale particles and structures like nanotubes are ‘passively’ incorporated into materials to enhance existing properties (as in the canonical examples of self-cleaning glass and lighter, stronger tennis rackets), before moving on to the eventual bottom-up assembly of inorganic and organic structures. This narrative, in which research here and now is represented as leading necessarily towards the evolution of new forms of “building” and “engineering” in the future bases itself squarely on the ‘Feynman vision’ of nanoscale control, and is represented, with sometimes very different and even conflicting interpretations of the route researchers need to take to realise this vision, by the writings of key figures within and outside the NNI:

As knowledge in nanoscience increases worldwide, there will likely be fundamental scientific advances. In turn, this will lead to dramatic changes in the ways materials, devices, and systems are understood and created. Innovative nanoscale properties and functions will be achieved through the control of matter at its building blocks: atom-by-atom, molecule-by-molecule, and nanostructure by nanostructure. ([70]: 1)

In assembler-based chemistry, nanomachines will bring molecules together to react only when and where they are wanted. Compared with biological chemistry, this strong control will enable assemblers to build structures of both greater complexity and generality. ([21]: 29)

In self-assembly, we rely on the importance of surface forces at the nanoscale to stick things together. We depend on the continuous random jiggling of Brownian motion to assemble the components and bring them together in the correct orientation for them to best fit together […] The benefits of using self-assembly are that, if we design and synthesize the right molecules, then all we have to do is to throw them together in a container to make our complex three-dimensional structure. ([42]: 90)

Whether the journey towards this vision is undertaken through techniques of mechanochemistry or self-assembly/biomimicry, it is typically seen as encompassing several stages that represent increasing levels of technical perfection, such as Roco’s four-stage characterisation [69]. The narrative leads us from the discovery of basic tools and techniques through to the higher levels of control required to produce stable nanostructures that replicate the successes of industrial manufacture at the nanoscale. In some cases it is linked to a general ‘convergence’ of technological advances occurring sometime during the next few decades which will represent a radical revolution in the human capacity for controlling the uncertainties associated with ‘natural’ phenomena like disease, aging, and scarcity of resources (e.g. [7, 71]). In this sense, a recurring teleological assumption may be traced in diverse nanotechnological imaginaries, which points to the common ground that exists between them and Drexler’s heavily-contested vision: “we have arrived at the stage where the assembler revolution is already contained in our present, making it inevitable” [53].

The structure of these narratives, it has been argued, relies on a reductive interpretation of processes of socio-historical change [53]. First, these processes are interpreted as being shaped primarily by advances in engineering techniques (technological determinism). Secondly, the possibility of such advances is interpreted as grounded in the fundamental nature of reality itself, which is assumed to be amenable to control based on specific techniques. It can be argued, therefore that, behind these narratives of technical progress are implicit ontological assumptions which define what nanotechnology is by what it can, or perhaps could do. These assumptions are contained within “constitutive metaphors” ([50]: 39–41) which provide a consistent basis for grasping and holding onto the invisible reality with which NSST concerns itself. These metaphors therefore help shape the technical possibilities of nanotechnology, by guiding research towards certain priorities.

For an example of such a metaphor, consider the imagery of “basic building blocks”, which is encountered again and again in contributions from neutral observers and critics and well as supporters of NSST (e.g. [23]: 6, [4]: 6, [80]: 46–7, [40]: 7). One reductionistic implication of this metaphor that has been articulated explicitly by NSST proponents is that the ability to isolate and manipulate the basic constituents of nature necessarily leads to “a better understanding of nature”, and that this knowledge brings with it the ability to re-order the configuration of these elements with a high degree of freedom and an increased chance of beneficial outcomes ([70]: 5–6). In its more extreme manifestations, this implication of the building blocks metaphor makes the nanotechnologist appear as a kind of architect who imposes a desired stable order upon matter, and in the process collapses any ontological distinction between inorganic matter and organic life ([45]: 61–3). A further implication of this image is that uncertainty tends to be viewed by scientists working on the control technologies that form much of cutting-edge NSST research as a subjective effect of the current limitations of techniques. Maria C. Powell suggests that scientists working on bottom-up assembly techniques who employ building blocks imagery in describing their work, tend to interpret the uncertainties they have to contend with as primarily being about “controlling and manipulating cellular processes or cellular interactions with nanomaterials, rather than uncertainties about cellular processes per se or how nanomaterials might interfere in these processes in harmful ways in organisms” (2007: 179–80).

If technique is understood as the controlled manipulation of objects, involving feedback effects which lead to gradual improvements in the level of control, then learning about uncertainty may tend to be undertaken from a position of comparative certainty regarding the relationships between past and future events. If techniques are thought of as being exercised within certain controllable parameters, the possible outcomes of action can be defined as a homogenous set of events that are related to past events ([18]: 578) by system variables that allow them to be defined as more or less desirable. From the knowledge of how less and more desirable outcomes vary in frequency over time, it is possible to estimate the probability of positive and negative outcomes in the future. In this way, an estimate of the risk of particular negative results occurring can be obtained, based on the degree of certainty with which statistical regularities can be drawn from the available data. Some events are not homogenous in the required sense. These are ones which cannot be adequately related to previous events through the system variables that are available (such as an event caused by another system impinging on the one being modelled). Such events appear more or less sui generis, unless an expanded frame of reference can be provided with an appropriate choice of variables. In the terms of the classic definition provided by the economist Frank Knight, such events are therefore uncertainties rather than risks ([51]: 233). By ‘uncertainty’ here is meant a possible event which cannot be expressed as a risk within the frame of reference used to describe a system. This definition therefore entails that, provided sufficient data is obtained, uncertainties can be in principle transformed into risks.

By defining the difference between risk and uncertainty in such a fashion, it is assumed that the assessment and management of uncertainties can be performed rationally and transparently. Among the practices employed to do this, policy tools such as risk-cost-benefit analysis (RCBA) have gained popularity [3] as ways of standardising and formalising assessment of the risks attendant on the use of particular techniques. As risk is commonly taken to be an objective attribute of an event, the use of methods of quantitative risk assessment (QRA) have therefore typically been separated within organisations from risk management, which is held to rest on value judgements that should not impinge on the risk assessment process (e.g. [16]).

As well as the assumption that future uncertainties can in principle be assigned a determinate and objective probability of occurring, QRA also rests on a further assumption, designed to maximize the homogeneity of the events that it studies. This defines the seriousness of negative events in terms of the likelihood of their occurrence multiplied by the severity of the damage that would most likely be done should they occur (expressed in terms of a financial measure, number of fatalities, or some other relevant standard). Taken together, these two fundamental assumptions ensure that risk assessment tends to interpret the meaning of uncertainty univocally as the statistical expectation value of an unwanted event ([33]: 3–4).

The technical progress narratives I have examined occur both in strategic visions of global nanotechnological futures [53] and in scientists’ own self-understandings [64]. What ties these narratives to the discourse of risk is the assumption that to understand technologies and their role in changing the world, they should be understood as bundles of techniques which produce control. Control represents an increase in both practical and in disinterested theoretical knowledge. The ubiquity of these assumptions about the essentially technical nature of technological evolution, together with that of the building-block ontological metaphors which help define relationships between technical interventions and natural processes is demonstrated by the way in which these assumptions and metaphors also inform many of the counter-narratives which have been introduced by those critical or sceptical of the potential applications of NSST. Many of these have been formulated in order to address political and/or ethical concerns which they claim have been overlooked by the dominant narratives of NSST-led ‘technological revolutions’.

Some of these base themselves squarely on calls for a precautionary approach to the risks of radical technologies, intended as a brake on the commercialization of the technology before a better characterisation of risks is available. In some of its publications, the ETC Group accepts the dominant characterisation of NSST as a set of techniques of superlative control:

Atomtech’s arrival is enormously significant because it gives us unprecedented potential to control and manipulate all matter—living and non-living. Atomtech is the great enabler—it offers access to a new realm, a molecular playing field where the building blocks for powerful technologies converge [23: 71].

This power, they argue, brings with it unknown hazards, in the face of which they invoke the precautionary principle, calling for a moratorium on the production of nanomaterials until the health and environmental risks associated with their use are better understood. The largely unregulated spread of a new technology is, on this view, morally wrong as it fails to respect the principles of good risk management. For the ETC Group, the control over matter which NSST may make possible cannot be free of the possibility of unintended consequences further down the line, once nanotechnologies are “in the wild”. The main benefits they foresee of the enhanced control offered by NSST would be for corporate interests, given the role played by technological innovation for economic growth. Any unexpected costs which may emerge will fall largely on the shoulders of wider society, and possibly on those who are least able to bear them. Given this concern with unintended consequences, the wider social risks of NSST development are thus a major concern for the Group, such as the potential nanotechnological applications may have for increasing disparities in economic power between the Global North and South [24].

Another counter-narrative has been advanced in reaction to the rhetoric around ‘convergence’ and ‘human improvement’ from within the non-utilitarian or communitarian wing of bioethics, informed by publications such as the President’s Council on Bioethics report Beyond Therapy [44]. This reading has adopted what Dupuy [22, 245] calls a “theological” perspective on the ontological implications of NSST, by interpreting the metaphors of nanoscale intervention and control as pointing towards the re-engineering of human nature, understood as an essential and given set of qualities which define human-ness [19, 287–288]. As a result, the risks of NSST are understood in terms of its potential to transgress transcendent moral constraints which limit the kinds of risk which should be taken. The application of technical mastery as an instrument of greater satisfaction of needs or preferences has to be reined in by an understanding of the limits which should be placed on such efforts in order to safeguard human dignity.

Both these counter-narratives tend to accept the technique-based definition of the nature of NSST. Their criticisms of the visions of progress embodied in the dominant narratives focus on factors which affect how the mastery that NSST may achieve will be applied, and what the wider risks of its application (in terms of environmental damage and social justice, or in terms of our ‘vocation’ as human beings) may be. In this regard, they modify, but do not overturn, the understandings of technological uncertainty which inform the dominant narratives.

The precautionary narrative accepts the authority of NSST’s definitions of itself and its technical potential, and also implicitly accepts the authority of established forms of risk knowledge over the interpretation of uncertainty [46]. It relies on the definition of unknown risks set out by extant official versions of the precautionary principle, i.e. risks which are currently unknown but which, in time, will be discovered and understood ([55]: 1800; [6]: 17), a definition which is entirely in line with Knight’s concepts of risk and uncertainty. The “theological” narrative, for its part, concerns itself less with issues of technical risk management and more with the ways in which the social meaning of the promised modes of nano-control could be considered transgressive of certain values prescribed by the fixed parameters of ‘human nature’.

What neither of these narratives make central to their objection to the dominant stories about NSST is whether the concepts of technique and bottom-up control are themselves sufficient to adequately understand the social meaning of a technology, and further, whether an understanding of uncertainty based on metaphors of technical control reflects and reproduces the dominance of some social relationships over others. The social constitution [8] of technologies is what makes them technologies (see e.g. [25]) rather than merely theoretical and practical knowledge of techniques and their products, isolated within a laboratory. To understand technologies as social is to define them as being constituted by a wider matrix of discourses (including symbolic representations and world-views) and practices (including techniques, norms, values and institutions) in relation to which they have been developed and are deployed ([22]: 241). As Hunt ([39]: 193) puts it, with respect to NSST: ‘[i]t is not as though first we shall develop nanotechnology and then decide how to use it’: the potential uses of NSST have already been moulded by the processes through which research priorities are shaped and different techniques are selected and developed. For example, the role of patent law in shaping basic NSST research has been recognised for some time, with much comment resulting on how each has exacerbated tendencies within the other. The result has been the early patenting of the basic ‘enabling’ components of NSST research (including some of its products, such as C60 nanotubes), which threatens to destabilize existing patent systems and restrict future research [52, 81]. This relationship between research and legal practice opens up the possibility that metaphors of building blocks and bottom-up control reinforce tendencies within the development of existing patent systems by encouraging researchers to treat successfully manipulated ‘blocks’ as intellectual property.

If this constitutive relationship between research on techniques and wider social processes is accepted, then our analytic focus shifts onto technologies as social phenomena. With this change of focus needs to come a substantially modified understanding of uncertainty. The technique interpretation promotes the view that uncertainty is, in essence, a temporary effect of technical imperfection, or in effect an epistemological problem. This is true of recent attempts to outline a general risk management framework for nanomaterials, which still tend to define the uncertainties surrounding NSST in terms of complex or uncertain risks whose nature will be more precisely defined as more knowledge is gained about relevant causal processes ([67]: 163–4). It is also possible, however, to view uncertainty as an element of the contingency which characterises the human condition, which is not the same thing as the theological counter-narrative’s concept of a set of intrinsic properties that make up the essence of what it is to be human. The condition of being human comprises the changing historical, cultural and natural constraints under which human beings make themselves and their world. Although these constraints change, what does not change is the finitude imposed by the social constitution of action, i.e. the fact that human beings act within the world under historically given conditions [22]. Uncertainty, on this view, necessarily and objectively accompanies finitude, rather than being an accidental side effect of our current state of knowledge. As such, it is strictly speaking more than just uncertainty; finitude implies a condition of radical contingency. If the ways in which humans experience the world are part of such a condition, then it is always the case that knowledge of apparently necessary processes can, with a shift of perspective and a reconsideration of the object of concern, lose its nomological character. For example, it might be discovered that knowledge of a set of natural processes is contingent upon conditions which were established prior to an investigation and which formed the basis of the investigation itself. Knowledge is inseparable from acts of selection and simplification.1

What sustains finitude is the gap between knowledge and the meaning of action, in the sense that action and its extended effects tend to exceed the compass of present knowledge, This means that the simple qualitative distinction between risk and uncertainty, or even a graded distinction between different “levels” of uncertainty (e.g. [82]) is not adequate to capture key conceptual distinctions and relationships between the ways in which the contingency of the future can insinuate itself into the present. In fact, the different varieties of contingency which the next section will examine can overlap ([85]: 116) within the gap between action and knowledge. What makes this observation regarding finitude particularly pertinent to our concern here with NSST is that, as some authors argue, certain technologies and social institutions are capable of greatly exacerbating this effect ([2, 75], 55, 105–6). Without placing finitude at the centre of any account of the relationship between risk and other forms of uncertainty, it will be difficult to address this tendency, which could be identified with what Beck calls organised irresponsibility [11].

Mapping Contingency

If we begin from the premise that the development of scientific techniques always takes place either in the service of the direct or the indirect pursuit of socially-defined ends, then the development of such techniques is technological in the sense defined above. Contingency, then, can be viewed either primarily from ‘inside’ the research processes that lead directly to the development of scientific techniques and technological products, or from ‘outside’ them. From inside, the shapes contingency takes have to do with the limits of scientific explanation and prediction with respect to natural processes. From outside, they are formed by social processes, the meaning of which is not determined prima facie for actors by knowledge of natural processes (Table 1).
Table 1

Types of contingency

1. Contingency as viewed from ‘inside’ research on techniques and technological products

2. Contingency as viewed from ‘outside’

a) Risk

a) Indeterminacy

b) Uncertainty

b) Trust

c) Ignorance or “non-knowledge”

c) Decision horizons

d) “Nescience”

d) Commitments and values

I now offer a characterisation of these different types, together with an outline of how their interrelationship could be understood.
  1. 1.

    Contingency as viewed from ‘inside’ research

  2. a)


As discussed in the previous section, future contingent events which may result from social activities are constructed in contemporary societies as manageable objects by being located within a homogenous population of events and then assigned probabilities. This understanding of uncertainty as quantifiable risk originated with thinkers like Pascal and Fermat in the 17th century, and the economic practices of merchants involved in overseas trade [14]. Quantitative risk assessment (QRA) draws on past data about the statistical distribution of particular kinds of events. It uses this data to estimate the probability of a given event in relation to other similar events and, by using either a single or multi-attribute system of values, strives to assess the seriousness of the likely consequences, should the event occur. In a single-attribute assessment, a single metric (typically, monetary value) is used to measure the possible size of losses. In multi-attribute approaches, different possible consequences are assigned different metrics ([61]: 70), so as to counter the criticism typically made of single-value approaches that the monetary values they assign to goods such as human health are entirely arbitrary [35, 84]. On this basis, the risk of different negative events can be analysed and the overall cost-benefit balance of the action assessed against that of any alternatives (including, in most cases, the option of doing nothing). QRA therefore offers a formal methodology, based on systematised past observations and on the current state of scientific knowledge in those areas relevant to particular assessment tasks. The legitimacy of QRA-based approaches therefore lies in an approach which, in its ideal form, would be based on “value-free” processes of assessment which could be applied across a variety of different situations. This understanding of QRA as a regulative ideal for risk analysis generally rests on the assumption that a separation between fact and value, i.e. between risk assessment and risk management, could exist (e.g. [16]). The results of QRA are envisaged as being supplied to decision-makers who then deliberate on which actions should be carried out and what measures taken to prevent or mitigate their likely negative consequences. In general, risk offers decision makers a framework in which the rationality of actions can be assessed on the basis of the degree of certainty available that they will produce a net benefit.
  1. b)


As noted in the previous section, not all possible events can be adequately dealt with within the formal framework of known risk. For example, in practice we may know the general parameters that are important for understanding how a given physical system works, without being able to assign definite probabilities to some events that may occur within it ([32, 34]: 44, [67]: 164). Whereas with risk a distribution of probabilities for specific outcomes is known, for situations that are uncertain in this specific sense this is not so. Uncertainty as ‘known unknowns’ tends to become a problem with increasing frequency as scientific knowledge is applied to the development of specific products, as the need to understand complex, singular situations replaces the experimental manipulation of processes within rigidly controlled laboratory environments ([43]: 201). Often uncertainties at this level and their possible effects on outcomes can be estimated and factored in to risk assessments, on the basis that they mark reasonably well-defined limitations on current knowledge ([85]: 114). In this sense, there is a degree of certainty attached to scientific uncertainty. It is defined by the boundaries of current knowledge of probable events that also mark out the limits of risk knowledge.
  1. c)



If uncertainty points to a set of incompletely known probabilities of events within a nevertheless relatively well-understood system, ignorance implies that such an understanding of the system is itself lacking. Importantly, ignorance can be viewed in two ways, based on a distinction which mirrors the one I have already made between technique and technology. If the feedback relation between scientific knowledge and action informed by it is understood as leading to the progressive improvement of techniques, then ignorance becomes implicitly defined as a condition produced by what remains outside the current boundaries of what science knows. This might be, for example, thought of as a body of facts which, once they are illuminated, will make the consequences of techniques more predictable. However, if the relation between scientific knowledge and action is viewed as part of a developing technology which is inextricably connected to wider social practices that shape the use of techniques and the development of scientific knowledge, then this positivistic view of ignorance is inadequate.

Instead of being defined a subjective condition which derives from those objective facts about the world which we have still to discover, ignorance now becomes a product of attempts to know more about the world, and as a concept which informs social practice, this version of ignorance differentiates what Ulrich Beck [11] has called reflexive modernity from a modernity which retains an unquestioned faith in a linear progress leading from ignorance to knowledge as more facts pass from the realm of the unknown to that of the known. Writers in the sociology of knowledge have suggested that ignorance is better conceived of as non-knowledge, which is produced alongside knowledge in two distinct ways. First, generating scientific knowledge not only has the effect of increasing our understanding of how limited our knowledge previously was and still is ([29]: 743). Its incorporation in technologies also leads to the creation of new objects whose potential for unpredictable interactions with complex systems ([75]: 111). Secondly, as scientific progress is not a process driven (or even determined in the last instance) by relations between ideas, it has to be understood how wider economic, political and social relations between business, government, and civil society create ignorance. At this point, the view of contingency from ‘inside’ research opens out onto the view from ‘outside’. The hope embodied by the regulative ideal represented by QRA—that risk knowledge can be made progressively more determinate—is undermined as a result. The appearance of objective knowledge separated from the processes which produce it becomes only an appearance, an effect of social prioritization. An example of when this becomes evident is when different priorities and methodological assumptions can produce ambiguity, i.e. opposed judgements about what outcomes can actually result from a given process [76].

The production of scientific knowledge comes about through complex social processes of prioritization through which certain problems and investigative practices are selected and other possible research foci ignored ([29]: 748). In other words, ignorance is an objective feature of the relationship between knowledge and reality, but not in the sense that we simply lack the full facts. Rather, scientific investigation tends to emphasize connections between aspects of reality in ways conditioned by a prioritization of problems that does not itself originate from within scientific research, and which is itself a generator of funding and with it, a determinant of what directions for research are viable. Andrew Barry gives an example of how this process operates in his discussion of air quality monitoring by remote sensing in Southwark in London, detailing how, amongst other non-scientific factors, UK and EU policy goals shape the selection and presentation of information, which has the result of creating “air quality” as a contested object of analysis ([9]: 155–9). This prior selection of priorities implicitly defines some possible objects of knowledge as salient and others as ones about which we do not need to know ([12]: 300).
  1. d)



All the modes of contingency discussed so far represent ones in which some form of knowledge can in principle be specified and can therefore serve as a guide to action. Alongside these varieties, however, sociology of knowledge has shown significant interest in recent years in ‘unknown unknowns’, that is, the irremediable ignorance which accompanies the production of knowledge [66]. This ‘nescience’ or, in Beck’s terminology, Nicht-Wissen-Können ([12]: 302), is the prerequisite for ‘total surprise’ ([29]: 751), and expresses our incapacity to predict all the consequences of our actions, no matter how far we are able to extend our understanding of determinate risks. Whereas uncertainty and ignorance can still offer some guidance to action, nescience reflects ignorance of which we are not conscious until an unexpected event has occurred. It refers to potential outcomes and especially unintended consequences which were not even imagined before the fact. These entirely unforeseen surprises can originate from even relatively banal interactions of familiar and well-understood processes, such as in the case Ian Hacking discusses of interference effects arising from the use of baffles inside cooling-tower chimneys for preventing the emission of ash ([31]: 148–150), as well as in situations when actions intervene in complex systems or generate effects which take years to fully emerge.

Crucially, all these modes of contingency typically interleave to create the different dimensions of uncertainty within which research is carried out. Some uncertainties involving more or less self-contained technical systems can be satisfactorily resolved without conflict between interested parties on the basis of risk analysis. But most often, there are difficult questions to be resolved by researchers about just what features of a situation can be defined as risks, uncertainties, or subject to ignorance. Conflicts can arise on the basis of these differing interpretations of the meaning of contingency. This is even more the case when the meaning of research is open to interpretation by other actors.
  1. 2.

    Contingency as viewed from ‘outside’ research

  2. a)



At this point, the difficulty of firmly drawing a distinction between putatively value-free risk assessment and value-driven risk management becomes apparent. The relationship between scientific investigation and wider social practices is, as noted above, complex even when the practice of scientific research itself is explored on its own. When the processes through which scientific work relates to policy making or technological innovation are examined, this complexity inevitably increases. The nature of the contemporary ‘contract’ between science and society, namely, the roles science is called on to play in informing policy and advancing innovation ([9], 173), places scientists in situations which are characterised by indeterminacy ([85]: 115–117), and which contribute to the production of ignorance (in the previously introduced sense of non-knowledge).

On the one hand, scientific investigation has to interpret pre-existing social priorities about how contingency is to be managed in order to properly understand its own remit. For example, scientific research which aims at establishing estimates of risk has to take a ‘slice’ through a particular social context in order to isolate what are felt to be important variables ([32]: 373). That this selection of what variables to study already reflects social priorities does not, however, often get reflected on at this stage. Brian Wynne gives the example of how, following the Chernobyl disaster, the persistence of radiocaesium in sheep living on the Cumbrian hills in the UK was found to contradict earlier scientific predictions about its environmental persistence ([85]: 120–122). The reason for this, Wynne argues, was that the measurements (dating from the 1960s) on which the earlier predictions were based had been guided by a particular set of social assumptions about the kind of risk-scenario that would be relevant in the case of environmental contamination by radiocaesium (i.e. a dose of radiation to an individual standing on the ground in a contaminated area). But after Chernobyl, a different exposure scenario (contamination via the food chain from grazing sheep to humans who ate them) became the focus of attention, and different parameters of the system under observation (the chemical mobility of radiocaesium as well as its physical disposition in the soil) were measured as a result. In this case, the re-focusing of policy interest on different modes of exposure served to render previous scientific research on risk obsolete. Changing social commitments revealed nescience and non-knowledge to have been unacknowledged limitations on previous scientific certainties.

On the other hand, after an assessment is carried out, a ‘downstream’ transformation of the evidence also occurs. For example, decisions on what actions to take implicate officials and managers who are responsible for implementing the results of these decisions ‘on the ground’ (as when risk-assessment-based regulations are used as the basis for implementing standards). At this point, the influence of social priorities which guide scientific investigation is modulated by social commitments which can selectively interpret knowledge, distort it by omission, or override it due to the pressure of sets of competing priorities, which may range from explicit value judgements about what kind of risks are economically or socially acceptable, to implicit cultural differences in the interpretation of risk analyses across the world in all the places a technology is put to use.

Indeterminacy therefore means that scientific knowledge cannot in principle contribute to decision-making the degree of certainty which policy makers have often desired from it.2 It opens up the framework in which decisions about risks are made to three other forms of social uncertainty.
  1. b)



First, we will consider the ‘downstream’ uncertainties which are implicated in the reception of expert testimony by decision makers and the public. Trust in expert evidence is vitally important where scientific research on risks is part of the management of contingency. Trust can however itself be a dimension of contingency: the question of who to trust for advice or to settle the evidential basis for action can be difficult to answer due to conflicts between experts, or to a lack of legitimate expertise given the extent of uncertainty and ignorance surrounding a problem.

Trust often depends on types of reasoning which cannot be reduced to the technical rationalities which prevail in risk assessment and in scientific investigation more generally ([26]: 57–58). Such forms of reasoning are arguably justifiable when, as is often the case, policymakers do not understand much of what experts claim to be certain about and cannot therefore judge whether these claims can be relied on or not ([34]: 45). Further, the emergence of non-knowledge, arising when different experts use different variables and investigative priorities in studying a single problem, can often lead to disagreement and explicit conflicts over what is relevant evidence ([32]: 380–381), as well as differing assessments of future outcomes which render the meaning of the problem increasingly ambiguous ([76], 310). Frequently, these conflicts themselves have an implicit or explicit political dimension, in the sense that it is often wider commitments that have led different experts to focus on particular aspects of a problem, thus creating conflicts which cannot be solved by referring to scientific data alone ([41]: 230; [59]). It has been argued that suspicions concerning the habitual assumptions of both industry and regulatory actors and how they may affect future management of risks and uncertainties have been contributory factors in some documented public perceptions of NSST [48, 54].

As a result, it may be hard to decide whether expert evidence should be trusted as a suitable medium for framing a set of contingencies, given that expert knowledge is itself produced through a set of largely indeterminate processes that are linked to the broader commitments experts, as members of society, hold about what is socially relevant and desirable [37, 74]. This leads us to the broader problem posed by the contingency of what matters in making decisions about action.
  1. c)

    Decision horizons


Questions concerning just what a decision to act or not to act is about are relevant both upstream and downstream of the points where scientific investigation linked to risk assessment and evaluation occurs. These questions concern which social priorities are most relevant to framing the decision making process, and also which commitments have already helped to construct the problem as it is understood. Often, such questions only emerge once a decision has already been taken, in public controversies where the values underlying an action and its supporting decision are criticised.

A ‘horizon’ of those wider values which are relevant to a decision has to be selected, implicitly or explicitly, in order to frame the terms of reference for a decision ([32]: 373–374). For example, consider a problem of toxic waste disposal. Should the goal be to manage risks in the locality in which the waste is disposed of? Should the goal be to formulate a broad industrial strategy to minimise waste production? Or should it be to promote a radical revision of the ways in which a society produces and consumes in the pursuit of economic growth? Upstream, the implicit decision horizon underlying a scientific investigation of risks (such as the priority of the health of a specific human population over long-term environmental damage) will also constrain the search for evidence.

Without settling on a specific horizon (a choice which will be constrained to some degree by professional responsibilities or the remit of an institution), the context for the decision will be entirely indeterminate. The choice of decision horizon may also be a factor in settling questions of trust, as some evidence can be ruled out on the basis of its irrelevance.
  1. d)

    Commitments and values

Finally, there is the potential indeterminacy of the ultimate values or ends which define these value horizons. In contemporary governance frameworks, these are typically framed to fit in with a broadly utilitarian view of the provision of goods and liberal-democratic assumptions about the limits of state power—the state as facilitator of markets and provider of the legal safeguards deemed necessary to minimise any harms they produce (see [79]: sec. 1.9). However, the legitimacy of these ends can itself become uncertain—for example, where the decision will affect individuals or groups whose sets of values may overall vary greatly from our own. This last has been thought to be particularly important where the possible consequences of action may affect future generations ([32]: 383–385).
  1. e)

    Relationships between modes of contingency

In the relationship between the ‘inside’ and ‘outside’ aspects of contingency, indeterminacy is therefore a key factor in how the meaning of contingent futures is interpreted and managed (see Fig. 1).
Fig. 1

Relationships between modes of contingency

In both cases, practices through which indeterminacy is shaped and made determinate operate with assumptions, forms of rationality and decision procedures which are not derived from scientific methodology. Research into and assessment of risk factors are themselves based on pre-formed priorities concerning the nature of socially-relevant harm, exposure scenarios and so on. After the research takes place, beyond the decisions taken about what should be done to manage risk, the effects of risk assessments affect wider social interpretations of contingency, through processes of media amplification or attenuation of risk, legal decisions, insurance practices etc. In both cases, the indeterminacy of future contingencies means that the interpretation of futures is contested based on assumptions about how to settle issues of trust, what decision horizons matter, and what ultimate values are relevant. This multi-dimensional nature of contingency means that to organise social action primarily on the basis of more or less quantifiable risks tends to emphasize the importance of one aspect of contingency (the possibility of directly harmful causal impacts on quantifiable variables such as incidence of health conditions, environmental or economic indicators) at the expense of others. As a result of basing action on assessments of risk, judgements about the rationality of action taken by regulators and managers tend to focus on particular ‘moments’ of decision and whether or not specific technical criteria have been applied. They will not extend to assess the institutions and wider, temporally-extended processes within which these decisions take place [60].

Broadly speaking, the adoption of a particular technical perspective to guide any investigation of a specific system requires a selection of what is to be studied and a specification of the relationships between its elements. This implies that any technical perspective always extends out from a background of social acts of prioritisation which embed techniques within technologies. Things are no different when contingency itself is constructed as an object of knowledge on the basis of certain techniques, such as risk analysis. These techniques already reflect certain wider values and practices, and thus bring with them ignorance, non-knowledge, nescience and indeterminacy even as they construct uncertainty as different from risk. The simplification which is necessary to produce knowledge in the first place is also the condition which makes surprises possible, once human beings act on the basis of the knowledge they have gleaned and blend the results of their actions with the complex systems they have sampled in their research.

Contingency as such is therefore rooted in finitude, the fact that the conditions under which we act and which we strive to understand the world are not, as Marx said, ultimately of our own choosing. As Jean-Pierre Dupuy notes [22], drawing on the work of Günther Anders and Hannah Arendt, this is an existential fact about the relation between human knowledge and action: knowledge is shaped by decisions, and vice versa, but there is no way to step outside this relationship. All production of knowledge implies a set of conditions under which it is made possible, but which are not necessarily themselves available for scrutiny alongside the knowledge that results. Thus it is not enough to argue that more scientific research is needed in order to resolve uncertainties, and that it is impossible to act until it is obtained. Rather, the circumstances in which human beings make their world are such that contingency is a perpetual accompaniment of knowledge ([22]: 246, 250). In other words, the background against which the different types of contingency come into focus is that of human finitude (see Fig. 1). Narratives of inevitable progress towards control, based on the unique possibilities of nanotechnology, thus represent a highly questionable interpretation of how futures are produced from present actions. Instead of opening a space for debate in which the limits of knowledge are placed at the forefront of discussion, such narratives (and the social contexts which they shape) tend to define uncertainty as what remains after the production of knowledge has been carried out.


The problem of uncertainty is most often posed as a matter of developing forms of social learning designed to remove or reduce it. In developing such practices, uncertainty no longer simply represents the ‘unknown’. It becomes an object of knowledge under different forms. As the history of risk management demonstrates, in specific historical contexts the methodologies which model the relationship between knower and known can become widely entrenched as instruments of governance, and as a result, normative expectations about how uncertainty should be responsibly managed gradually become accepted and unquestioned ([86]: 460–1).

So far, there is evidence to suggest that the uncertainties surrounding NSST tend to be framed—by technology strategists, technologists, and policymakers, as well as critics of actual and potential NSST applications—primarily in terms of risk and problems of control. The processes of social learning which such framings demand are ones which target uncertainty as a subjective artefact of the current state of technical knowledge, and demand more research to establish more firmly what ‘the risks’ are, an increasingly universal demand in the face of a slow international regulatory response from governments (e.g. [3, 58]). This tends to focus concern on causal impacts of NSST products, and away from the indeterminate processes which lead e.g. to the prioritization of some research goals and not others. It is perhaps notable that the outcomes of recent exercises in extensive public engagement on the possibilities of NSST in the UK have shown that, although people often see the need for ‘more research’ to specify particular health and environmental risks, other important areas of concern have to do with the indeterminacies surrounding the technology, particularly the values and ends it serves, and the full extent of uncertainty and ignorance surrounding the processes in which it seeks to intervene (e.g. [28]: 42–3; [62]: 205–6). What I have argued here is that the contingencies surrounding NSST should be approached, in the first instance, by focusing on technologies and not techniques, and on indeterminacy rather than risk. A key step forward here would be to address the complex provenance of scientific knowledge, to which Funtowicz and Ravetz’s concept of “pedigree” is intended as a contribution [27]. A further step would be to review the procedural and substantive ethical bases upon which current understandings of stakeholder engagement rest, an ongoing task with which various commentators within the philosophy of technology are engaged (e.g. [49, 77]).

Risk is undoubtedly a significant part of the landscape of uncertainty (as Fig. 1 above indicates), but to frame it as the most significant is to ignore a number of other vital dimensions of uncertainty, and indeed to invert the relationship between risk and these other dimensions—almost, in effect, treating a map as if it were the territory. Even the idiom of ‘ethical, legal and social impacts’ (ELSI) represents a risk-related framing of the issues opened up by indeterminacy, one which allows wider issues of contingency to be shunted down the list of concerns thanks to the difficulty of addressing them through technical risk research. On the view I have presented here, the engagement exercise participants who talk about the indeterminacies of ‘control’ and ‘progress’ are expressing a rational approach to uncertainty, and with it, an appreciation of the deeper political and ethical significance of uncertainty that demands further investigation.


The concept of contingency articulated here has perhaps its closest philosophical antecedent in Hegel’s philosophy. See Hegel ([36]: 542-45) and the discussion in Žižek ([88]: 153-56).


On how the BSE crisis demonstrated the unrealistic nature of policy-makers’ expectations of scientific advice, see Adam [1]: 164–91.


Copyright information

© Springer Science+Business Media B.V. 2009