Living with Uncertainty: the Limits of ‘Risk Thinking’ Chris Groves1 Introduction Responding to climate change demands of us that we learn how to handle responsibly something we have created – emissions of greenhouse gases (GHGs) – whose capacity to have unpredictable causal effects on complex systems may endure for a span of time greater than that of recorded human history. They extend the ‘timeprint’ (Adam and Groves 2007) of human action beyond the reach of forms of knowledge which, through observation, measurement and naturalistic explanation, have the aim of predicting the future through the past. Since the Renaissance, the gradual coming-to-dominance of such forms of knowledge in the study of nature has been accompanied by experimental attempts to extend the model of explanation they offer to the study of societies. The representation of such knowledge as value-free expertise in managing social action for the public good, exercised either directly through social reform carried out by government bureaucracies or indirectly through efficient investment and production by private businesses, has created the image of the wise manager as the symbol of a peculiarly modern rationality (Macintyre, 1981: 82). The public or private manager plays a role within contemporary societies which is played by shamans, prophets or priests within societies not dominated by Weberian bureaucracies. S/he is held to have access to forms of knowledge which allow her to effectively mediate between past, present and future. His/her scientifically grounded expertise in guiding social change against a background of uncertainty serves as the ground of his/her legitimacy. If such expertise derives from a naturalistic form of investigation and explanation, then it also depends upon a particular image of the future. The growth of empirical natural science was inseparable from the emergence of this image of an empty future. This is the future seen more or less exclusively as an object of mathematical calculation, without theological or ethical significance that can be articulated without reference to such forms of knowledge (Adam & Groves, 2007: 69-75). With the increasing authority of natural science, the empty future gradually became a fundamental presupposition of social practices, particularly in the case of institutionalised risk management. Evolving initially alongside early commercial capitalism, these practices became more formalised with the historical process of industrialisation, eventually reaching high levels of mathematical sophistication alongside the refinement of probability theory (Hacking 1990; Bernstein 1996). In the 20th century, through the spread of what I shall call ‘risk thinking’, particular approaches to the management of social uncertainty became definitive of management expertise. Their acceptance as a normative standard for decision making transforms the social future as such into a calculable object (Rose, 1999), with the result that the ethical value of a policy or decision begins increasingly to be derived from how far it is judged as having the potential for filling an otherwise empty future with more economic growth, more welfare, more generalised happiness. A problem with the normative framework of risk thinking emerges, however, when the management of risk has to deal with products (e.g. GHGs) which extend our ‘timeprint’ beyond the reach of naturalistic, predictive forms of knowledge. Here, the ethical legitimacy of risk thinking breaks down 1
ESRC Centre for Business Relationships, Accountability, Sustainability and Society (BRASS), Cardiff University, 55 Park Place, Cardiff, CF10 3AT. Email:
[email protected]
in the face of the relationship between present and future. In fact, reliance on risk thinking in this characteristic contemporary situation can be interpreted as an abdication of responsibility for the future. Risk Thinking and Control What is risk thinking? A loose system of concepts and practices employed for managing uncertainty by private and public institutions, which becomes more tightly consolidated when it is embedded in legal instruments (such as regulations and standards). The role played by metaphors in organising the relationships between different concepts and practices within risk thinking is crucial (cf. Lakoff and Johnson 1980; Klamer and Leonard 1994). One such metaphor might be uncertainty is risk; another such, knowledge is control. The growing acceptance of these metaphors as basic assumptions about reality is one result of a new contract between scientific research institutions and society which emerged out of the historical process of industrialisation, and the promotion by public and private managers of technology as an instrument for steering social change by creating objectively measurable forms of progress (such as growth in GDP). Through the wisely managed use of technology, uncertainty could itself be conquered. At the beginning of the eighteenth century, Jean-Baptiste Vico gave the postulate of the “new science” (1725) a celebrated formulation: Verum et factum convertuntur (“The true and the made are convertible”). This means that we can have rational knowledge only about that of which we are the cause, about that which we ourselves have produced. (Dupuy 2007, 256) What we have made, we understand; what we understand, we control. Vico’s maxim implies that the behaviour of that which is controlled need no longer be subject to uncertainty. Consequently, the use of technology to drive social change must therefore be an efficient way of maximising benefits and avoiding harms. The same forms of knowledge which led to the initial creation of technologies would, in the hands of expert scientific advisors, provide managers with the information they need to use their own expertise in managing the social impacts of technologies most effectively. We can therefore add another, still more basic metaphor to risk thinking’s inventory of assumptions: making is knowledge. If we gloss Vico’s maxim further by considering the relation between the metaphors making is knowledge and knowledge is control, it becomes plain how these together relate to the identification of uncertainty with risk. Taken together, they imply that making is control, i.e. the imposition of order or form upon an unordered matter. In this way, a qualitative difference is imparted to matter. But this derives from an incremental, quantitative increase in order, which finally reaches the point where the product of our exertions harmonises with our purposes. As we transform matter, we begin to understand more about the laws that govern it. Consequently, we seek to extend our control over both the functional properties of our products, and over the uncertainties that may surround their typical modes of operation. A steam engine, for example, might be designed both for efficient production of mechanical force, and for safety in the case of pressure vessel failure. Knowledge about risk is therefore produced when the uncertainties which surround the operation of an object themselves become objects of the same kinds of knowledge which inform the processes through which the object is manufactured. The mathematically-assisted management of natural forces during manufacture is complemented by mathematical estimations of how likely interactions between these forces are to produce undesirable consequences.
If knowledge and certainty are therefore associated with control, ignorance and uncertainty are necessarily associated with a lack of control. Uncertainty, ignorance and lacking control are subjective phenomena associated with a failure to thoroughly investigate the external world, whereas knowledge, certainty and control stem from grasping the objective nature of things through the desire to impose order upon them. The mathematical tabulation of risk is therefore a mapping of objective features of the world, as opposed to mere uncertainty, which is a feature of the contingent limitations of our knowledge (Knight 1921, 233). As control and knowledge advance, the dark pool of ignorance with which we are surrounded is gradually illuminated, and uncertainty is transformed into determinate risk. In this way, it is clear that the metaphors of knowledge and control we have listed organise the relationship between knowledge and ignorance in a particular way, dramatising it as a progressive and incremental conquest of the latter by the former. Similarly, they support the construction of the third metaphor we listed above: uncertainty is risk. By suggesting that when we make, we reveal the workings of the world, they imply that by making things we also make things manageable. By intervening in the world, a simple lack of certainty about the future is transformed into a well-ordered ranking of probable outcomes. The increasing use of technology as a means of steering social change, which emerges from industrialisation and accelerates after WWII, encourages the extension of the three metaphors we have examined to processes of social change themselves. This, in turn, stimulates the spread of risk thinking. Professional practices of risk analysis were originally associated with relatively isolated technical systems, such as chemical plants or aircraft systems. In such systems, the technical processes and parameters are well-defined, and the reliability of separate components is testable or amenable to actuarial in-service analysis. […] [T]his original cradle of risk analysis allowed its authors to build in assumptions of well defined and deterministic processes (Wynne 1992, 113) However, since WWII, ‘risk’ has come to denote practices of uncertainty management which extend over a domain much more diffuse, complex and tangled than the operational context of a discrete technical system. The result is a union between risk analysis concepts as used in engineering, actuarial practices drawn from insurance and finance, and concepts from welfare economics such as that of expected utility. The amalgamation of these concepts and practices has consolidated the concept of risk as a means of understanding and managing the uncertainties associated with different policy choices. As risk comes to cover not only the uncertainties inherent to the immediate operational context of technical systems, but the uncertainty of the social future itself, risk thinking becomes an essential tool of self-legitimation for private and public sector managers alike. The interpretation of uncertainty in terms of risk establishes a standard for rational decision making. In the next section, we will briefly examine the application of risk thinking to climate change, and explore how the identification of uncertainty with risk leads to an abdication of responsibility for the full timeprint of our actions. The Politics of Uncertainty Why does it matter that society increasingly interprets uncertainty as risk? In short, because this interpretation of uncertainty is a political and ethical choice. The selection of some aspects of uncertainty for attention rather than others “distributes both the power to control relationships and freedom of action unevenly between [society’s] members” (Marris 1996, 1).
Anthropogenic global warming (AGW) caused by increasing GHG emissions is a theory which seeks to explain certain long-range climatic phenomena, specifically an increase in the Earth’s average globally-measured temperature over the last 60 or so years. What the future will look like on the basis of the generally-accepted models of the relation between GHG emissions and average temperatures is uncertain. Instead of a single answer, there is a multiplicity of potential scenarios based on variations in GHG output and assumptions about climate sensitivity. These scenarios suggest that the likely increase in temperature will be between about 1.5°C and 4.5°C, depending on variations in the relevant system variables. Higher increases are not ruled out, but the probabilities of such increases are not treated as reliably definable. What impacts a substantial increase in global average temperatures might have on different societies is a matter of huge uncertainty. In response to the growing body of evidence on AGW amassed by the IPCC, social managers have undertaken the task of transforming this uncertainty into well-defined risks, in order to school politicians in the nature of their responsibility to present and future generations. By examining how this task has been carried out, we can discern the hidden ethical and political assumptions of risk thinking. Given the timeprint of AGW, these assumptions present problems which are not just adventitious ones that can be solved by more research. Here, the belief that ignorance can gradually be transformed into knowledge, and a lack of control into control, must falter. The problems risk thinking faces are endemic to it as a system of thought and practice built on the three metaphors we explored in the last section. The highly influential Stern Review (Stern 2006), widely taken to be the most authoritative AGW policy response to date, represents a comprehensive and sometimes admirably self-critical attempt to understand the uncertainties surrounding the phenomenon. However, it also arguably presents an excellent example of how risk thinking misrepresents the problems involved when the timeprint of human products necessitates taking a long-term view of future-oriented responsibility. Stern’s conclusions, very briefly summarised, are that the risks associated with continuing increases in the emission of GHGs are of an annual loss of around 5-20% of global GDP, every year. If action is taken to limit emissions to between 450 and 550ppm of CO2 equivalent, then, it is argued, the process of AGW can be controlled and these risks avoided for an estimated cost of 1% global GDP each year (Stern 2006, vi-vii). The arguments by which these conclusions have been reached have been trenchantly criticised, however, precisely on the grounds that their extension of concepts of risk to the relationship between climate and social change involves reductive and therefore ethically and politically illegitimate interpretations of the nature of uncertainty. Baer and Spash (2008) note that uncertainty has weak and strong forms. Weak uncertainty is of the kind discussed in the previous section. Subjective in nature, it lies outside current knowledge of determinate risks probabilities. Nonetheless, it can still be factored into a calculation. Risk estimates are based on past knowledge of a system’s behaviour, and this knowledge can help define for us factors which we do not yet know but which can potentially be characterised, or at least accounted for. Strong uncertainty, on the other hand, derives from factors such as ‘unknown unknowns’ or the unpredictability of social action. Because AGW offers the possibility of unique and possibly extreme or catastrophic changes in climate and social systems which are (a) outside past human experience and (b) which may occur at any point within the timeprint of GHG emissions (an indefinitely long future), then the uncertainties that surround it are of the strong rather than weak variety. Assuming that the range of possibilities can be best dealt with by relating them to incrementally accumulated knowledge of past events is therefore itself problematic. For Baer and Spash, Stern’s decision to adopt
this assumption results in a speciously accurate quantification of risk in the face of insuperable uncertainties, one which argues arbitrarily for the 450-550ppm limit, when lower limits could be justified on different assumptions. Crucially, the key assumptions on which Stern’s reduction of strong to weak uncertainty is based are ethical rather than methodological. These form the basis for his interpretation of the characteristic rationality of risk thinking. These assumptions, common across most social applications of risk, integrate concepts and practices borrowed from technical risk analysis and welfare economics, and represent a reflection and interpretation of pre-existing social priorities. In this sense, the objectivity which is attributed to formal risk analyses of, say, a chemical plant or aircraft system cannot be attributed to the metaphorical extension of risk to more complex systems, whether social or nonsocial. To come to conclusions about social impacts requires acts of interpretation on the part of the risk manager that are, at bottom, judgements about what matters, and are therefore ethical and political. The chief example in the Stern Review cited by Baer and Spash is the employment of costbenefit analysis (CBA) as a means of estimating the risks of different AGW scenarios. CBA represents a way of extrapolating from past knowledge to future estimates of how the costs and benefits of a particular policy will add up. Stern’s use of CBA leads to conclusions about the effects of different levels of control over GHG emissions on aggregated global expected utility. Stern’s final judgement about how much GHG emissions can justifiably be curbed by is therefore based on a weighing up of costs and benefits that ultimately places a loss of consumption by some in the present in the balance against the future deaths of other, poorer individuals. The controversy over using CBA in this context is precisely about such issues as assuming we can legitimately trade lost lives for consumer goods – a dilemma which symbolizes debates over commensurability [...] In the end, the numbers produced by the Stern Review are only meaningful if one accepts that the prospective human deaths (plus extinction of species and other losses) due to human induced climate change can be defensible converted into equivalent amounts of consumption today (Baer and Spash 2008: 25). Stern’s assumption is that the economics of AGW, in allowing a quantitative estimate of risks attendant on action now to be weighed against those of allowing business as usual, can set the terms of ethical and political debate about what actions are justifiable. But this is to get things the wrong way around: assuming that, for example, the loss of lives can be equated to some equivalent loss of consumption is a key assumption behind Stern’s use of CBA, but this is politically and ethically contestable, particularly given that the loss of life may well be unequally distributed between rich and poor countries into the distant future. By interpreting uncertainties in terms of economic concepts that are central to risk thinking, Stern assumes that we can be responsible for our timeprint by considering how much we are prepared to pay here and now to control GHG emissions. Whether this is a correct interpretation of what this kind of responsibility might really require is not, however, a question which he considers, or indeed which can be considered within the limitations of risk thinking. Conclusion The extension of risk thinking to wider social and non-social systems rests on dubious assumptions, as the example of AGW shows. Risk is a concept which evolved within a particular technological context. But the contexts within which it is increasingly applied are ones in which the rationalities of control upon which it rests are politically and ethically contestable. The rationality of risk thinking is supported, as we saw, by metaphors of knowledge which represent ignorance and uncertainty as a simple lack of knowledge and certainty. But sociologists have argued that uncertainty is not a simple lack. This is because knowledge is inevitably based on a limited selection from the possible objects of
knowledge, and an equally limited exploration of some of their interrelationships under artificial conditions. Ignorance is therefore actively created alongside knowledge, and has therefore been characterised as ‘non-knowledge’ (Beck 1996; Gross 2007). The selection of what to study is constrained by pre-existing social interpretations of ‘what matters’. For example, the development of a fossil-fuel economy in which the main sources of investment are geared to obtaining maximal shortterm profits constrains technological development to pursue the exploitation of hydrocarbons as cheap energy inputs, rather than allowing for widespread experimentation with alternative energy sources and research into the systemic impacts of GHG production. Ignorance, uncertainty and a lack of control are not adventitious consequences of the current state of our knowledge: they are necessary consequences of the finitude which is part of being human (Dupuy 2007), our inability to comprehend all the conditions of our limited perspective on the world. If risk thinking is based on an extension of a naturalistic model of knowledge which is one more element of this, our finitude, how do we deal with the vast uncertainties that surround the timeprints of our actions and products? In confronting this question, we have to recognise that the empty, calculable future with which natural science operates is a potentially paralysing image which threatens to foreclose possibilities for ethical reflection once it is extended to encompass the social future. We have to ask ourselves what rationalities, and what images of the future, remain to us once the limitations of the empty future envisaged by natural science and bureaucratic management have been exposed.
References Macintyre, A. (1981). After Virtue. London: Duckworth. Adam, B. and C. Groves (2007). Future Matters: Action, Knowledge, Ethics, Leiden, Brill. Baer, P. and C. L. Spash (2008). Cost-benefit analysis of climate change: Stern revisited, CSIRO Working Paper Series 2008-07. Online at: http://www.csiro.au/files/files/pkec.pdf. Beck, U. (1996). Wissen oder Nicht-Wissen? Zwei Pespektiven reflexiver Modernisierung. In: Reflexive Modernisierung. U. Beck, A. Giddens and S. Lash. Frankfurt a. M., Suhrkamp: 289-315. Bernstein, P. (1996). Against the Gods: The Remarkable Story of Risk, Chichester, Wiley. Dupuy, J. P. (2007). "Some pitfalls in the philosophical foundations of nanoethics." Journal of Medicine and Philosophy 32(3): 237-261. Gross, M. (2007). The unknown in process - Dynamic connections of ignorance, non-knowledge and related concepts. Current Sociology 55: 742-759. Hacking, I. (1990). The taming of chance, Cambridge, Cambridge University Press. Klamer, A. and T. C. Leonard (1994). So what's an economic metaphor? In: Natural images in economic thought. P. Mirowski. Cambridge, Cambridge University Press: 20-54. Knight, F. H. (1921). Risk, Uncertainty and Profit, Boston, MA, Houghton Mifflin. Lakoff, G. and M. Johnson (1980). Metaphors we live by, Chicago, London, University of Chicago Press.
Marris, P. (1996). The politics of uncertainty: attachment in private and public life, London; New York, Routledge. Rose, N. (1999). The Powers of Freedom: Reframing Political Thought, Cambridge: Cambridge University Press. Stern, N. (2006). The Stern review: the economics of climate change, London, HM Treasury. Wynne, B. (1992). Uncertainty and Environmental Learning - Reconceiving Science and Policy in the Preventive Paradigm. Global Environmental Change-Human and Policy Dimensions 2(2): 111-127.
