How to Shape a Better Future? Epistemic Difficulties for Ethical Assessment and Anticipatory...

How to Shape a Better Future? Epistemic Difficultiesfor Ethical Assessment and Anticipatory Governanceof Emerging Technologies

Brent Daniel Mittelstadt & Bernd Carsten Stahl &N. Ben Fairweather

Accepted: 4 March 2015 /Published online: 14 March 2015# Springer Science+Business Media Dordrecht 2015

Abstract Empirical research into the ethics of emerging technologies, often involvingforesight studies, technology assessment or application of the precautionary principle,raises significant epistemological challenges by failing to explain the relative episte-mic status of contentious normative claims about future states. This weakness meansthat it is unclear why the conclusions reached by these approaches should beconsidered valid, for example in anticipatory ethical assessment or governance ofemerging technologies. This paper explains and responds to this problem by proposingan account of how the epistemic status of uncertain normative claims can beestablished in ethical and political discourses based on Jürgen Habermas’ discourseethics. To better understand the nature of the problem, the relationship betweennorms, facts and the future is explored in light of potential meta-ethical fallaciesfaced in the field of empirical ethics. Weaknesses of current approaches to anticipa-tory ethical assessment and governance are then explored, including the PrecautionaryPrinciple and Technology Assessment. We argue that the epistemic status of uncertainnormative claims can be understood within Habermas’ approach to political discourse,which requires ‘translation’ of uncertain claims to be comprehensible to other stake-holders in discourse. Translation thus provides a way to allow for uncertain normativeclaims to be considered alongside other types of validity claims in discourse. Thepaper contributes a conceptual account of the epistemic status of uncertain normativeclaims in discourse and begins to develop a ‘methodology of translation’ which canbe further developed for approaches to research and ethical assessment supportinganticipatory evidence-based policy, governance and system design.

Keywords Emerging technologies . Empirical ethics . Evidence-based policy . Technologyassessment . Epistemology. Habermas . Anticipatory governance

Ethic Theory Moral Prac (2015) 18:1027–1047DOI 10.1007/s10677-015-9582-8

B. D. Mittelstadt (*)University of Oxford, Oxford, UKe-mail: [email protected]

B. C. Stahl : N. B. FairweatherDe Montfort University, Leicester, UK

1 Introduction

It has long been established that technologies can cause ethical problems (Dusek 2006). Thereis broad agreement that an early recognition of these issues may provide avenues to addressthem through anticipatory policy and regulation. A core issue is that a tension exists betweenthe empirical and the normative dimensions of dealing with future and emerging technologies.This is the heart of the Collingridge dilemma (1980) which states that it is impossible to knowwith certainty the consequences of an emerging technology at an early stage when it would becomparatively simple to change the technology’s trajectory. Once the technology is moreestablished and it becomes clearer what its social and ethical consequences are goingto be, it becomes increasingly difficult to affect its outcomes and social context.Despite the fundamental problems posed by the Collingridge dilemma, there continueto be attempts to better understand future technologies and to predict the ethical issues theyare likely to raise with a view to addressing them early through anticipatory policy, regulationand design choices.

There exists a rich history of attempts to develop epistemologically robust ways ofunderstanding how emerging technologies will affect the future, for example in future studiesor technology foresight (Cuhls 2003; Georgiou 2008). Such approaches often attempt toprovide an empirical basis for policy makers due to a growing demand for evidence-basedpolicy (Banks 2009). Despite this rich field of research, there is still no agreement on some ofthe epistemological underpinnings and the question of how our understanding of possiblefutures can or should be translated into practical (political) activities that will allow for thedesirable shaping of future technologies.

At the heart of the conceptual disagreement is a core epistemological problem: proposedanticipatory actions intended to influence the shape of the future are based on ‘uncertainnormative claims’ which are inherently uncertain (because they describe what the future willbe) and prescriptive (because they describe a good or desirable future). Uncertain normativeclaims, or claims about how the future will and should be, have epistemic and normativecomponents which combined to blur their epistemic status. Such claims are (1) epistemolog-ically uncertain because they describe the future, which is logically unknowable, and (2)normative because they work on some conception of a good or desirable future, whilepotentially making claims about how it can be brought about in practice through anticipatoryactions. As this suggests, the label ‘uncertain normative claims’ can be applied to the reasonsand evidence used to support taking particular anticipatory actions in response to potentialconsequences of (for example) emerging technologies. Considered together, the epistemolog-ical uncertainty and normativity of uncertain normative claims indicate that the justificationunderlying anticipatory actions such as precautionary policy or modifications to system designlacks clear epistemic status. Without a clear account of how these components change thejustificatory power or ‘epistemic status’ of uncertain normative claims in comparison to (forexample) empirical ‘facts’ or legislative mandates, the support offered for anticipatory actionsis weakened.

If anticipatory actions taken in response to potential consequences of emerging technolo-gies are based on normative claims about the future which are inherently uncertain, then suchactions require two types of justification: first by demonstrating that the predicted shape of thefuture (how the future will be) is reasonable or convincing, and second by showing that theproposed action(s) to change the future (e.g. a normative claim about how the future should be)is desirable or acceptable. The former may be provided through foresight or future studies (forexample), while the latter requires ethical argumentation or discourse to justify one’s positionand underlying values against which a good future, and the actions required to reach it, are

1028 B.D. Mittelstadt et al.

defined. An account of the epistemic status of the uncertain normative claims, or claims howthe future will be and should be, is therefore required to allow such claims to be evaluated andweighed in the ethical and political discourses which shape emerging technologies, resulting injustified anticipatory actions.

Practically speaking, uncertain normative claims can be seen in evidence andargumentation provided to policy-makers and others which attempts to influencefuture regulation. Significant research is conducted into public attitudes and emergingtechnological developments to provide evidence for such anticipatory governance. Inaddressing the key epistemic problem outlined above, it is therefore necessary to considerhow the epistemic status of uncertain normative claims can be connected to the methods ofresearch and public consultation throughwhich they arise. We suggest in this paper that, despitethe inherent uncertainty of the claims and research aiming to provide (evidence for) action-oriented advice to policy and decision-makers to address or pre-empt future problems, reason-ably validity claims (and therefore, anticipatory actions) can be made while uncertainty stillprevails.

While the primary purpose of this paper is to argue for the existence of problematicepistemic uncertainty underlying anticipatory actions such as evidence-based policy, and tosuggest that this issue needs to be addressed by researchers providing evidence (and thusuncertain normative claims) for evidence-based policy-making and similar anticipatory ac-tions, we also describe an account of the epistemic status of uncertain normative claims basedaround Habermas’ discourse ethics.1 The approach we suggest is to treat future-orientednormative claims arising from empirical ethics research in the same way as other contentiousvalidity claims. Drawing on Habermas’ recent works (2008, 2011), we argue that normativeclaims about the future can be treated analogously to religious claims in political discourse,according to which contentious validity claims must be ‘translated’ so their evidence base iscomprehensible to others in the discourse. The epistemological and normative components ofuncertain normative claims must be ‘disentangled’ before the claim can be reconstructed or‘translated’ and thus made comprehensible to participants in an ethical (or political) discourse.This allows for inherently uncertain claims encountered in empirical ethics researchconcerning possible futures (e.g. emerging technologies and their consequences) to be incor-porated in broader political discussion of emerging technologies and public policy. However,while we can initial identify two types of components (epistemic and normative) in justifica-tion anticipatory actions, separating them in practice when deciding which action to take maynot be so simple.

To demonstrate how this approach can be applied in practice we then present an exampleresearch project and explain how uncertain normative claims were encountered and ‘translat-ed’ to clarify their epistemic and normative components. While not a full-fledged ‘methodol-ogy of translation’, the pilot work undertaken in the project described may provideavenues for further conceptual development of our approach to addressing this keyepistemic problem facing responses to possible futures brought about by emergingtechnologies. This leads to the conclusion which explains how normative empiricalresearch can build on our conceptual ideas in order to play a more prominent and appropriaterole in public discussions and policy advice.

1 We are not making the claim that the only valid philosophical approach to address the epistemic difficulties with‘evidence’ for future-oriented policy and governance is Habermas’ discourse ethics, but rather that Habermas’works provide one viable approach to explain the epistemic status of normative claims about the future, which isrequired to justify the usage of such claims as evidence in evidence-based policy and similar anticipatory actions.

How to Shape a Better Future? 1029

2 Norms, Facts and Emerging Technologies

Building on the key epistemological difficulties discussed above, the core problem of thispaper is the relationship between empirical research and anticipatory action with regards tofuture, emerging, and therefore uncertain technological developments. In order to clarify ourcontribution to understanding this relationship, a brief exploration into prior relevant dis-courses is required. What follows is a review of established positions on the relationshipbetween facts and norms, description and prescription. This leads to a discussion of ethicsunder conditions of uncertainty which is followed by an analysis of current attempts tocoherently incorporate normative empirical data into ethical analysis. Finally, a reconciliationof the epistemic difficulties of uncertainty with the need for anticipatory normative evidence-based action is offered.

2.1 The Relationship of Facts and Norms

To be justifiable, anticipatory decisions require a legitimate source of morality, as determinedby relationships between current and ideal practice, or facts and norms. Conceptualising theappropriate interaction between facts and norms is a core problem of ethics and by no meansnovel. Moral philosophy has developed a plethora of approaches that aim to provide anunambiguous source of morality. Well known examples include the Kantian basis ofnormativity in reason (Kant 1827, 1870), the utilitarian idea of deducing normativity fromutility (Mill 2002) or the hope that normativity can reliably be derived from intuition orsentiments (Smith 2007). This paper does not offer the space to do this long and complexdiscussion any justice. The main reason for mentioning the problem of the source of moralityis that it relates directly to the question of the relationship between normativity and empiricalresearch. This is another well-established problem, aspects of which have been discussedunder headings such as the Bis-ought problem^, BHume’s Guillotine^ and the Bnaturalisticfallacy.^

The core of this discussion goes back to Hume’s (2004) view that ‘ought’ cannot bededuced directly from ‘is’, that prescription needs to be independent of description, and that‘objective moral facts’ are therefore impossible. Moore (1993) asserted a related albeit distinctposition, saying that ‘good’ is not a natural property, meaning any attempt to define it requiressubstituting other properties of something that is good. Movements from ‘is’ to ‘ought’ inwhich the ‘is’ statement is of the type ‘such and such is good’ are therefore legitimate forMoore, as ‘good’ can describe a real property of things and events. Despite these differences,both arguments rely upon a fundamental conceptual distinction between ‘is’ and ‘ought’ whichrequires any normative conclusion to have a normative premise.

The distinction has been broadly upheld in empirical work on technology, for example innormative approaches such as value-sensitive design (Friedman et al. 2008). At the same time,the conceptual segregation between ‘is’ and ‘ought’ has met with resistance. Magnani (2007)concurs with Searle (1964) that the intellectual tradition strictly distinguishing between ‘is’ and‘ought’ is Bbizarre.^ He argues that the truth of a statement implies a prescription to believe it.Similarly, Floridi (2010) points to the tradition of classical thinkers from Plato to Spinoza whosaw the roots of ethics in being. He then bases his own conception of information ethics on amoral worth of information which arises from its being. A further example of such a naturalistposition that deduces ‘ought’ from ‘is’ was provided by Jonas’s (1984) Bimperative ofresponsibility^ where he argues that there is a new categorical imperative that we must notrisk human survival in light of the technical possibility of eradicating humanity. He postulatesa sort of meta-responsibility which requires the possibility of responsibility by safeguarding


human existence. A further figure of argument is that natural processes can inform ournormative intuitions. We can, for example, learn about ourselves and by extension about theorigins of our moral duties and obligations by observing the natural world and naturalprocesses, such as evolution (Mumford 1996). Relying on such observations for normativejustification necessarily includes logical steps that can be described as committing the natu-ralistic fallacy by those accepting Hume’s position, if no further premises or evidence aregiven. Observations can nevertheless provide further context and understanding of moralintuitions. A final and important aspect of a naturalistic approach to ‘is’ and ‘ought’ is thereliance on religious figures in moral thought. Adherents of religion tend to deduce normativestatements from their authoritative source and they are often related to beliefs about religiousfacts. The existence of a personal god may lead to the acceptance of normative prescriptionsthat emanate from this god. While this can easily be seen in many interpretations of Abrahamicreligion, a similar merging of ‘is’ and ‘ought’ can be observed elsewhere (Hongladarom 2005).

The purpose of this brief elaboration was not to provide comprehensive coverage of theproblem area (cf. Black 1964), but to set the scene for an exploration of the way in whichmoral philosophers, ethicists and social scientists work empirically, including their integrationof description and prescription in ethical analysis which can produce advice or evidence foranticipatory governance of emerging technologies.

2.2 Facts and Norms in Empirical Ethical Analysis

The conceptual distinction between description and prescription presents a practical problemfor researchers intending to use empirical information describing Bhow the world is^ to suggestBhow the world should be.^ However, an examination of the normative content of empiricaldata and its treatment in the last decade suggests that the importance of the problem may beoverstated. Historically, empirical data has had a supportive yet detached relationship withethics. Social scientists and ethicists often work separately on a single project, with the formerresponsible for gathering empirical data and the latter tasked with analysis and reachingnormative conclusions (Molewijk et al. 2003). This division of work itself stems fromaffirmation of the fact-value gap, according to which the descriptive world remains the domainof scientists, while the prescriptive world belongs to ethicists (Molewijk et al. 2003). Undersuch a separation ethicists have been limited to using descriptive empirical data (1) to applyethical theory in the context of a policy or action (Molewijk et al. 2003; McMillan 2008); (2) toassess the validity of assumptions upon which moral theories are based (De Vries and Gordijn2009; Leget et al. 2009); (3) to gain insight into social practices for the identification ofrelevant cases and stakeholders in ethical deliberation (van Hooren et al. 2008). Each of theseexamples upholds the conceptual distinction between description and prescription.

In contrast to prior practice, the last decade has seen the emergence of iterative integrationof empirical data in ethical analysis (Musschenga 2005; De Vries and Gordijn 2009). Thismovement, termed Bempirical ethics,^ has arisen from the recognition that empirical data canbe usefully employed in ethical analysis to build and translate ethical theory into Bmiddle-range principles^, refine ethically questionable practices and reach contextually-sensitivenormative conclusions for practical purposes including policy guidance (Birnbacher 1999;Musschenga 2005; Leget et al. 2009). Empirical ethics is best understood as a mentalityguiding the use of empirical data in ethical analysis. Proponents are far from unified in terms ofphilosophical background and methodology, yet they tend to share three basic assumptions: (1)studying the moral beliefs of individuals in a practice yields meaningful normative empiricaldata that should be a starting point for ethics (Borry et al. 2004; Baldwin 2008; De Vries andGordijn 2009); (2) descriptive and normative approaches are inherently complementary; and


(3) empirical ethics cannot rely on context alone for its determination of morality, but mustincorporate both empirical data and ethical theory (Borry et al. 2004; De Vries and Gordijn2009). Empirical ethicists therefore attribute value to the moral beliefs of individuals involvedin a practice (or Bmoral practitioners^), because the practice provides them with unique Bmoralexperiences^ (Musschenga 2005). This attribution of value is coherent with (but not limited to)the ontological and epistemological commitments of philosophical positions involving dis-course (e.g. discourse ethics, hermeneutics), which implicitly support normative empirical dataas a legitimate source of morality and starting point for ethical analysis. Accordingly, empiricalethics affirms that valid normative empirical data can be collected, challenged and refined byengaging with the experiences of individuals involved in a practice, or ‘moral practitioners’.The distinction between description and prescription in the gathered data may collapse inpractice (De Vries and Gordijn 2009) through iterative data collection, analysis and dialoguewith moral practitioners.

Debate over the conceptual separation of description and prescription will continue unabat-ed, but it should not prevent further exploration of the epistemic status and treatment ofnormative claims (or data) in ethical analysis. While empirical ethics has contributed to thisdebate within research into current practices, it has not yet turned its attention to the uniqueepistemic challenges presented by future practices and emerging technological developments.Evidence-based policy which responds to emerging technological challenges requires alegitimate source of morality for justification. Without an account of the epistemic status ofuncertain normative claims, the epistemic status and thus justificatory power of some of theevidence base for anticipatory governance cannot be established, complicating assessment andweighing of contradictory uncertain normative claims. To address this problem, a newconceptualisation of the relationship between description and prescription, or facts and norms,under conditions of uncertainty is required.

3 Ethics, Uncertainty and Emerging Technologies

Before a reconciliation of facts, norms and the future can be offered, the challenges ofuncertainty must be understood. Building on the discussion above, practices for which thesechallenges are relevant can be explored through consideration of current approaches to ethicalassessment of emerging technologies, of which the Precautionary Principle and TechnologyAssessment are perhaps the best known. In understanding the challenges, it is important tonote the difference between uncertainty and risk, although both concepts are related. Withuncertainty the probabilities of possible outcomes are unknown (Knight 1921), whereas withrisk Bthe probability distribution… is known or predictable^ (Ahmed and Skogh 2006, p183).In looking at technologies still under development, researchers are often faced with uncer-tainties for which even the minimal data required for the maths of imprecise probability isunavailable. With this said, the paradigmatic case considered here is a technology that issufficiently developed for meaningful discourse to be possible about the nature of thetechnology and its initial uses, but where there is still uncertainty about its future implications.Under this type of uncertainty of future technologies, one can ask on which grounds decisionsand policies can be made and justified.

In considering the challenges of uncertainty we are attempting to map the epistemic statusof ‘uncertain normative claims’. Uncertain normative claims arise in discourses as contentiousarguments with epistemic and normative components which describe the appropriate responseto future social and ethical impacts of emerging technologies, or what ‘should’ be done in thefuture. These claims are uncertain in several senses. Epistemologically, such claims are


uncertain regarding what should be considered knowledge regarding the inherently unpredict-able future, but also in terms of what should be done on the basis of this knowledge, the latterof which is also a normative question. As the future is inherently uncertain, the claims are notepistemologically equivalent with claims based upon existing social practices (cf.Widdershoven et al. 2009) or empirically observable phenomena (cf. Popper 1959).2 In termsof normativity, such claims are uncertain in terms of what a good future is, but also whatactions should be taken to reach this future. Normativity refers to any declaration withprescriptive or evaluative content, including statements of moral evaluation (such as Bthistechnology is good^) or prescription (such as Bone should do …^). If uncertain normativeclaims are to ground anticipatory action, a clear account of their epistemic status is requiredwhich acknowledges this dual epistemic and normative uncertainty. The need for such anaccount can be seen in difficulties faced by existing decision-making and assessment proce-dures guiding anticipatory actions towards emerging technologies.

3.1 Evidence-Based Decisions under Conditions of Uncertainty

The difficulties of anticipating and responding to the future have historically been addressedthrough the creation of decision rules to guide anticipatory action, which direct decision-making in absence of reliable empirical or statistical data regarding (for example) risks andimpacts. A notable early decision rule was the ‘rule of insufficient reason’ (Laplace rule)otherwise known as Bthe principle of indifference^ (Keynes 1921)—that when we are toounsure about the probability of possible outcomes Bwe should treat them as if they wereequally likely^ (Sinn 1980; Goodin 1983). This has been shown to be inappropriate in verymany circumstances, often inconclusive (Keynes 1921) since it is often arbitrary how wedivide options up into the events between which we are ‘indifferent’, and sometimesBintuitively absurd^ (Roncaglia 2009). Subsequent research has advanced on decision rulesat varying levels of uncertainty in a number of ways, resulting in a Bvast literature^ (Ahmedand Skogh 2006) including, for example the Bdiversification theorem^, which explains why itmay be wise to spread investments over a number of assets when the returns are uncertain andKeynes’ theory of liquidity preference, explaining how agents act to keep options open in theface of unforeseen developments (Roncaglia 2009).

The uptake in research reflects the on-going debate about decision rules under conditions ofuncertainty, largely provoked by debates about policies for new technologies, and especiallyenvironmental policies during the late 20th and early 21st centuries. Against this backgroundthe question of how to assign epistemic status to uncertain normative claims becomes urgent.Opponents of regulation have time and again argued for a position that has been characterised(Wagner 2003) as saying that regulation should not occur because there is Bnot enough scienceto justify protective regulation.^

In response proponents of regulation no longer rely entirely on trying to provide scientificproof that will satisfy their opponents. Sometimes this is because Bstandards for the requisiteevidence are never articulated^ (Wagner 2003) and—thus—never met to the satisfaction ofopponents of regulation. Even without scepticism towards the tactics of opponents of regulation,proponents may have realised that requiring scientific proof of harmful relationships to justifyregulation is inappropriate. There are two fundamental reasons for this. The first is the epistemicstatus of science: scientific theories are, fundamentally, provisional (awaiting falsification)

2 How validity derives from scientific observation is contentious question. We adopt Popper’s critical rationalistposition, in which all scientific knowledge is provisional and held up to the standard of falsification. Certainty,then, refers to the falsifiability of a scientific hypothesis or theory.


(Popper 1959), so conclusive evidence to justify regulation would forever remain out of reach.The second is because of the conservatism of science: in identifying relationships science doesnot work on the basis of balance of probabilities, but rather seeks to Bminimise ‘false positives’^,even though in doing so its Bprocedures increase the chance of ‘false negatives’, that is, failing toassert that there is a relationship when there is a relationship^ (John 2010).

Rather than always seeking to provide scientific proof that will satisfy opponents, propo-nents of environmental regulation have looked for arguments that regulation should beginbefore there is certainty about the harmful effects of the processes they want regulated; in otherwords, they seek decisions under conditions of uncertainty.

3.1.1 The Precautionary Principle

Taking a leading position among these arguments is the Precautionary Principle (PP), whichcame to global prominence with the Rio Declaration of 1992 (UNCED 1992). Although acommon definition is lacking (Gardiner 2006), the core content of the PP is found in theWingspread Statement (1998) about environmental impacts of human activity:

BWhen an activity raises threats of harm to human health or the environment, precau-tionary measures should be taken even if some cause and effect relationships are notfully established scientifically.^

From its creation the PP has had a severely (but justifiably) poor philosophical reputation(Hughes 2006; Gardiner 2006), although recent adaptations have sought to improve itsstanding. The principle’s reputation comes largely from the Bparadox of precaution^ (a termapparently coined, but not philosophically analysed, by Morris (2000)—the term Bparadox^had previously been applied by Freestone (1991) to the precautionary principle in a differentway). The Bparadox of precaution^ arises if the precautionary principle is interpreted as sayingthat we must avoid harm by eschewing risky, but potentially beneficial technologies. Underthis interpretation harm arises since the benefits of adopting the technologies are foregone withBfor example… lost opportunities to prevent disease and death^, resulting in a Bprecautionaryprinciple [that] would instruct us to refrain from implementing itself^ (Hughes 2006). Beyondthis, even enthusiastic advocates have been willing to accept it is Bill-defined and imperfectlytranslatable into codes of conduct^ (Jordan and O’Riordan 1999).

Work since 2006 (cf. Gardiner 2006; Hourdequin 2007; John 2010), however, has producedmore philosophically robust formulations, which may give valuable guides to action in certaincircumstances. Gardiner’s (2006) work turns on criteria for the employment of the PP. A keyinsight from Gardiner is that the PP applies in conditions of uncertainty, but not whenignorance exists of the range of outcomes over which there is probabilistic uncertainty.

Unfortunately, the PP lacks a mechanism for identifying and normatively assessing poten-tial outcomes. Such an Bidentification mechanism,^ guided perhaps by normative empiricalresearch, would help identify possible outcomes under conditions of uncertainty. Approachesto the ethical assessment of emerging technologies require such a mechanism which involvescross-referencing Bdescriptions of the technology…with ethical values and principles^(Brey 2011). Identification mechanisms therefore not only identify a list of possibleethical outcomes, but give normative content to these outcomes by locating them withinexisting ethical theories and concepts. The interplay between facts and norms can be said toexist in this process. Therefore, without such a mechanism, the PP is of little use inevidence-based policy making, and does not explore the relationship between facts andnorms in future-oriented ethical assessment.


3.1.2 Technology Assessment

Descriptive and prescriptive information relating to the potential outcomes of emerging technol-ogies is required to use the PP in decision-making. Such information can be gathered in empiricalethics research, in which normative claims are made about emerging technologies. Of theapproaches capable of collecting, analysing and applying such information under conditions ofuncertainty, Technology Assessment (TA) is perhaps the most widely used. TA is a field whichstudies and evaluates the interaction between new technologies and the environment, industryand society (Tran and Daim 2008). Similarities are found with other plausible approaches forresponding to emerging technologies including participative design (Mumford 2003), value-sensitive design (Friedman et al. 2008) and other socio-technical approaches (Sandberg 1985).Studies engaging in TA often work on Bknown or potential applications^ of emerging technol-ogies as well as future and technological forecasting studies (Georghiu 2008; Cagnin et al. 2008;Brey 2011). Crucially, TA includes an Bidentification mechanism,^ a normative interest in futuretechnologies, and a range of empirical methods to pursue this interest.

The field emerged in response to experiences of undesirable or unintentional side effects ofemerging technologies (Grunwald 2009). TA comprises a family of approaches that aim tocombine empirical research on likely consequences of technologies with normative insights.Its various flavours, including participative TA (Joss and Bellucci 2002; van Eijndhoven andvan Est 2002) and constructive TA (Genus and Coles 2005), employ a broad array of methodswhich include participation from a variety of stakeholders, most notably laypersons asrepresentatives of future users of the technologies in question.

By anticipating and responding to the challenges presented by emerging technologiesduring the development and early deployment cycles, TA studies aim to guide the develop-ment and regulation of technology in more desirable directions (Brey 2011). Grunwald (2009)identifies four challenges which TA responds to:

Bthat of integrating at an early stage in decision-making processes any availableknowledge on the side effects, that of supporting the evaluation of the value oftechnologies and their impact, that of elaborating strategies to deal with the knowledgeuncertainties that inevitably arise, and that of contributing to the constructive solving ofsocietal conflicts on technology and problems concerning technological legitimisation.^

TA is therefore united as a field by its emphasis on the production and evaluation ofknowledge concerning social, economic and environmental impacts for the purpose of pro-viding recommendations to steer the social response to emerging technologies.

3.2 Problems of Normative Evaluation in TA and PP

TA studies often refer to the Precautionary Principle in making recommendations for socialresponse to the uncertainties of emerging technologies (Grunwald 2009). This situation createsa problem for TA as most versions of the PP lack guidance on how to assess and weighcompeting uncertain normative claims. The reliance on the PP may be a conscious decision—anticipatory actions must surely be sensitive to the context in which the technology will beused, which will influence the relative value of uncertain normative claims. Still, accepting thislimitation of the PP ignores a common problem of any approach to the assessment of futuretechnologies. Even if normative assessments give importance to context (which we believethey should), there must be criteria by which the relative value of (uncertain) normative claimsis determined. While these criteria may change according to context, culture, religion, ethical


viewpoint etc., the need for an evaluative mechanism does not go away. Balancing competingclaims without knowledge of their epistemic status is akin to judging the scientific merit of acontrolled trial without an understanding of how controls contribute to the scientific method—the key piece of knowledge for determining the value of uncertain normative claims, theirepistemic status, is missing.

3.3 The Importance of Epistemology in Anticipatory Decision-Making

We are now at the point where we can propose an account of the epistemic status of uncertainnormative claims. To review, emerging technologies can raise ethical issues that policy makerswould like to address proactively. Being situated in the future gives such technologies uncertainoutcomes. Despite this, empirical research is still conducted to determine their possible ethicaland social outcomes. However, as a result of the epistemic and normative uncertainty of claimsmade about how the future will and should be, such normative empirical research requires adifferent type of truth claim compared with objectivist research in the natural sciences, whichbases truth claims validity measured in terms of falsifiability (Popper 1959). While decisionprinciples such as PP can give some indication of appropriate evaluative mechanisms, theseonly take effect once stakeholders in decision-making agree upon a shared view of facts andnorms. TA and related approaches set out to provide empirical input into such policy develop-ment and decision mechanisms. However, the question of the epistemological evaluation of theoutcomes of TA research remains open, which precludes a shared understanding of facts andnorms on which assessment and weighing of uncertain normative claims may occur. How can adecision maker who aims to make evidence-based policy interpret the validity of uncertainnormative claims about emerging technologies provided by researchers? We suggest in thefollowing section that this can be done by requiring the ‘translation’ of such empirically-basedclaims prior to their consideration in ethical and political discourses.

4 A Reconciliation of Norms, Facts and the Future: Discourse Ethicsand the Interpretation of Uncertain Validity Claims

The relationship between ‘is’ and ‘ought’, the way in which empirical ethics deals with thisrelationship, and the broader context of ethics under conditions of empirical uncertainty havebeen considered thus far. The purpose of considering these topics is to develop a soundtheoretical basis that can be used to reconcile empirical research and normative aspects offuture and uncertain technical developments. For this purpose we now need to show that thedifferent aspects under consideration can be combined in a theoretically satisfactory manner toprovide a coherent account of the epistemic status of normative empirical claims underconditions of uncertainty. There are, no doubt, numerous ways of providing such theoreticallysound conceptual foundations. We argue that, in terms of epistemic status, uncertain statementsconcerning normative aspects of emerging technologies can be treated analogous to religiousstatements which, according to Habermas (2008, 2011), should be treated as valid andlegitimate interventions in public discourses. This is not to say that uncertain normative claimsand religious statements share similar foundations (e.g. faith, superstition), but rather that bothpossess inherent uncertainty which must affect how they are treated in discourse.3

3 Certainty refers here to the possibility of falsification of the claim and its evidence base. Uncertain normativeclaims are characterised by reliance upon descriptions of the future which are inherently uncertain, whereasreligious claims rely upon faith or belief beyond empirically observable phenomena.


One appropriate way of achieving the aim of providing the necessary theoretical underpin-nings for normative research on emerging technologies is to make use of the existing body ofwork on discourse ethics. The rationale for this choice is that discourse ethics is a well-established ethical approach that deals with the relationship between normative and empiricalstatements. Discourse ethics furthermore has an established track record of being applied to(human interaction with) technology (Mingers and Walsham 2010). Finally, recent develop-ments in the broader political discourse theory as put forward by Habermas allow for theincorporation of statements that lack the epistemological certainty that can be expected instatements about empirical observations of existing phenomena. In order to develop thisposition, discourse theory and discourse ethics must be briefly introduced to demonstrate theirusefulness with regards to scientific and technical developments. It must be shown howuncertain normative claims can be accommodated in discourse.

4.1 Discourse Ethics

The term ‘discourse ethics’ was originally coined by Karl-Otto Apel and was adopted anddeveloped by Jürgen Habermas. Space limitations preclude a detailed and in-depth discussion ofdiscourse ethics; however, a brief overview is provided which will allow us to demonstrate thatdiscourse ethics provides a theoretical foundation of the problem addressed this paper. For thepurposes of this argument we will concentrate on the Habermasian version of discourse ethics.Discourse ethics must be understood in the broader context of Habermas’ theory of communi-cative action (TCA) (1981). According to the TCA human beings as social entities need tocommunicate in order to survive and prosper. Humans have a range of possibilities of commu-nicating, with communicative action being the best or most highly developed.Whenever humanbeings communicate, a set of validity claims arise: truth (Wahrheit), rightness (Richtigkeit) andauthenticity (Wahrhaftigkeit). Communicative action requires the speaker to engage in adiscourse whenever any of these validity claims are queried. This implies a willingness toengage with the interlocutor, to take her seriously and to be willing to change one’s positions inthe light of that argument (Habermas 1997). Such discourses may (or perhaps, should) arise in avariety of discussions and processes concerning the ethics, governance or design of emergingtechnologies, including pre-emptive policy-making and regulation in which the PrecautionaryPrinciple may otherwise be applied, participatory design processes involving stakeholders of thetechnology in question (e.g. Reuzel et al. 2001; Joss and Bellucci 2002; Friedman et al. 2008),and deliberation over specific proposed uses. Concerning the latter, such deliberation couldoccur for example in discussions between clinicians, health administrators and patients consid-ering commissioning emerging wearable health monitors for deployment among patients withCOPD (e.g. Pantelopoulos and Bourbakis 2010; Department of Health 2011).

This brief basis of the TCA is important for the current paper because it indicates that thereis no fundamental distinction between normative and descriptive statements. Any statementcan raise validity claims relating to truth, normative rightness and the speaker’s authenticity.All validity claims are subject to discursive questioning.

Habermas’ discourse ethics incorporates these principles of the TCA and uses them todevelop an explicit ethical theory (1991, 1992). Discourse ethics is expressed in two basicprinciples, the discourse principle and the universality principle. The discourse principle statesthat only those norms can claim to be valid that meet (or could meet) with the approval of allaffected in their capacity as participants in a practical discourse. The universality principle goesbeyond the acceptability by the affected and states that the consequences and side effectsarising from the general adherence of a norm have to be acceptable for all involved stake-holders and, ideally, for everybody.


Discourse ethics can be seen as a successor of Kantian deontology that incorporates theKantian principle of universalizability and overcomes the limitations of the categoricalimperative which is aimed at the individual. It is important to understand that discourse ethicsis one aspect of a larger theoretical edifice and needs to be interpreted in this context.Habermas’ work covers political and legal theory which are inextricably linked with ethicalquestions (1998). It is therefore not surprising that applications of Habermasian thought can befound in areas where technology is used for purposes of communication in the public sphere(Lyytinen and Hirschheim 1988; Heng and De Moor 2003). While much work usingHabermasian ideas uses his earlier and more sociological writings (Klein and Huynh 2004),it is important to note that recently Mingers and Walsham (2010) have made a convincingargument that discourse ethics proper has the potential to help us understand normative aspectsof technology.

Discourse ethics can thus be argued to overcome the limitations of the ‘is-ought problem’and provides a theoretical position that allows empirical ethics research. The discourse anduniversality principles provide criteria for evaluating norms which allow for normative as wellas empirical interventions. It requires practical discourses which allow the voicing of empiricalobservations as well as ethical positions.

4.2 Uncertain Normative Claims in Discourse

This leaves the final aspect at the core of this paper, namely the problem of uncertainnormative claims. According to the TCA any validity claim can be queried and then needsto be defended using good arguments in a discourse. This raises the question of how toevaluate and defend statements about the future that are fundamentally epistemologically andnormatively uncertain.

Within the framework of discourse ethics and the TCA at least two strategies will bepossible. One can use a methodological argument that defends statements with reference to themethodologies that were employed in arriving at the statements. A statement about anexpected social or ethical consequence arising from an emerging technology could thus bedefended by explaining the methodological approach that led to it. This is a rational way ofdealing with the uncertainty of the future but it has the disadvantage that it can turn discoursesabout the ethical aspects of emerging technologies into discourses on future-oriented researchmethods.

We believe that recent work by Habermas points to a better way of incorporating uncertainnormative claims into ethical-political discourses and anticipatory decision-making. In termsof normativity, uncertain normative claims are not fundamentally different from other norma-tive claims made in a discourse; both make equivalent validity claims open to questioning indiscourse while being based on moral values, ethical theories, social knowledge or othersources. Therefore, additional work is not required within discourse ethics to address thenormative components of uncertain normative claims. However, the epistemic components ofuncertain normative claims require further treatment to justify their inclusion in publicdiscourses. This goal can be accomplished by treating uncertain normative claims as analogousto religious beliefs. The relationship between the two is that they both occupy similarepistemological territory. Both types of claims rely upon or describe states of existence thatare logically unknowable—metaphysics for the former, the future for the latter. As a result,neither can be falsified or tested through scientific means.

Habermas has discussed the role of religion in political discourse in some depth since thebeginning of the current millennium (2008, 2011). This focus was the result of the continuingprevalence of religious interventions in political discourses and the recognition that these


interventions must be taken seriously because they can be strong motivators of political actionand the basis of legitimate democratic positions. Habermas’ question is how a post-metaphysical state, or a state that is not based on religious convictions or traditions, can reactto and incorporate religious positions. He argues that such a post-metaphysical state needs tobe secular and neutral towards religious convictions, in order to facilitate the peaceful co-existence of different, possibly contradictory religious positions. Religious citizens in a secularstate have a ‘duty to translate’ their religious claims into a secular or rational form if theyexpect their claims to be considered in discourse. At the same time, non-religious citizens havethe duty to accept that religious positions can have a rational basis which may be valid indiscourse, and have a right to be heard. There is thus the possibility of a discourse despite thefact that some interlocutors do not share the basic premises of the interventions of others. Onecan interpret this view as the implementation of the principles of the TCA in cases ofprincipled disagreement on the validity claims in statements.

This principle can be extended to cover uncertain normative claims arising from empiricalwork on future technologies. Such statements do not have the same epistemological status asstatements about empirically observable phenomena. At the same time, they carry values andmeaning attached by participants such that they may fruitfully contribute to ethical-politicaldiscourses that shape the development of such technologies. What is required from thedifferent participants is (a) the ability and willingness to translate these uncertain normativeclaims into practically relevant statements and (b) the willingness by interlocutors who doubtthe validity of such claims to listen to the translation in the attempt to use discourse to come toa better shared view of the social reality in which the discourse participants find themselves.

4.2.1 Translation of Uncertain Claims

The process of translation involves the reconstruction of unsupported and (perhaps) unjustifiedmoral beliefs about future states into legitimate ethical claims.4 Moral beliefs about futurestates are vulnerable to propaganda, misinterpretation of evidence and ignorance. Accordingly,it would be foolish to apply a kind of Laplace rule and view all moral beliefs as equally validwithout further analysis. We suggest that two steps are necessary to translate such beliefs intolegitimate ethical claims that can be subjected to discourse. First, the justificatory evidencebase upon which uncertain normative claims are constructed must be understood as transfor-mational experiences, flaws in reasoning or false statements can be identified (and perhapsengaged in discourse). Second, the Bnormative truth content^ of uncertain normative claimsmust be understood. To unpack this concept, it is helpful to review Habermas’ position on theutility of religious beliefs within a secular political society. Much as religious beliefs may relyupon adherence to questionable comprehensive doctrines, moral beliefs and normative claimsmay be founded upon indefensible or subjective premises. However, both types of beliefs maycontain legitimate conclusions or ‘normative truths’, albeit by reliance on questionablefoundations. The process of translation thus identifies the normative truth of questionablebeliefs and translates it into a defensible claim by (in the case of religious beliefs) translating itinto secular terms or (in the case of moral beliefs) relating it to ethical concepts and theory.This allows the different interlocutors to agree or disagree on the content of the claims on thebasis of mutual understanding.

4 A distinction must be made here between the concept of ‘translation’ recently proposed by Habermas (2008,2011) which is an action to be undertaken in public discourse, and ‘translation’ as used in Science andTechnology Studies, for example originating in the work of Latour (1987). Here we refer solely to the conceptas used by Habermas for purposes of establishing a rational discourse.


Conceived as such, the purpose of translation is not to convince other stakeholders of thevalidity of an uncertain normative claim, but rather to reveal the ‘meaning’ of the claim anddisentangle its epistemic and normative components for further discourse. Translation plays acritical role in ethical assessment under conditions of uncertainty clarifying the epistemic statusof uncertain normative claims by helping discard the metaphysical or uncertain elements of aclaim which initially blur its epistemic status. In practice, this requires converting the claiminto terms or a frame of reference understood by or shared with the other stakeholders; forinstance, translating religious claims between the metaphysical frameworks of different reli-gions, or into purely secular terms for other citizens in a secular society. For uncertainnormative claims about the future, translation may require identifying the principles orprescriptive claims which have been applied to the future context, or the characteristics of afuture context which give ‘weight’ to the uncertain claim. The translation of statements fromone context to another is thus the key to considering uncertain normative claims aboutemerging technologies as valid in evidence-based policy making. This raises a number offollow-on questions, such as: What constitutes a successful translation? Who needs to translatewhat and to which degree? Are translations reversible? How can a discourse deal withcompeting translations? The approach we suggest here thus requires the development of atheory or ‘methodology of translation’ with relevant evaluative criteria (perhaps based in theTCA).

4.3 Translation in the Ethical Assessment of Emerging Technologies

At present, a ‘methodology of translation’ does not exist. While the primary aim here is toshow that Habermasian translation is a plausible approach to comprehend the epistemicdifficulties of incorporating uncertain normative claims into anticipatory evidence-basedpolicy and the ethical evaluation of emerging technologies, demonstrating its practical utilitymay further this cause. If our account is successful, it follows that such discourses may beevaluated in terms of the translation of uncertain normative claims, requiring future develop-ment of a translation methodology, which goes beyond the scope of this paper. However, initialcontributions to such a methodology have been made in a doctoral research project recentlycompleted by the lead author (see: Mittelstadt 2013).

In this project, translation was carried out by breaking the claims made by respondents in aninterview-based empirical study into components. The premises, evidence and conclusionswhich constitute the uncertain claim were identified during translation to assess which piecesof a claim which rely upon uncertainty, facilitating questioning of particular validity claimsrather than the uncertain claim as a whole. In this way the normative truth content of uncertainclaims was separated from the uncertain framework in which it resides. It is the act ofseparation which constitutes translation—a claim is broken down into its constitutive partswhich can be understood by participants for further assessment in discourse. Translation thusallows for uncertain claims to be questioned in two different ways in discourse: first, are theuncertain descriptions of the future plausible given current empirical evidence and indicationsof the shape of the future?5 Second, are the normative position which rely upon the uncertaindescriptions of the future themselves acceptable to stakeholders, apart from the uncertainelements of the claim?

5 Criteria to determine credible predictions of the future will vary among stakeholders in discourse and betweendisciplines. The determination of acceptable criteria in this regard is a separate issue from establishing the needfor and process of translation.


In a sense, claims are turned into ‘if-then’ statements: if a certain state of affairs occurs inthe future, then (credible) prescriptive statement X is applicable. By separating claims into two(or more) constitutive parts, the validity of each can be considered separately in discourse.Discourse can then reveal incomplete, poorly supported or otherwise deficient components ofthe uncertain normative claim, while refining and strengthening components with a credibleepistemic or normative foundation. The importance here is on translation as a process in whichuncertain claims are separated into premises for further assessment, rather than as a set ofuniversal criteria by which uncertain claims are proven certain or given relative normativeweight to suggest which action should be taken. Participants bring into discourse variouscriteria for questioning and accepting the truth of a statement. Translation is the process whichallows for these diverse criteria to be applied to uncertain claims in discourses such asevidence-based policy-making, by separating the normative truth content of the claim fromits uncertain framework.6

4.3.1 Translation in Practice

An example of translation from the aforementioned project may help clarify how the processcan potentially be carried out in practice. Uncertain normative claims regarding the ethicalimplications of emerging technologies can be seen to contain three overlapping types ofcontent, each of which can be questioned in discourse. Consider the statement made by adaughter caring for her mother with dementia when asked about GPS tracking: BI wouldn’tfeel comfortable putting a GPS tracking chip under the skin of my mother in the future. Iwould feel like I was spying on her. I know she would hate it because she is a very privateperson.^ In this example, the following three types of claims are evident:

1. Characteristics of the Future—The technological characteristics, capacities or uses ofthe technology in the future the respondent believes will occur, e.g. BGPS tracking chipunder the skin.^ These are descriptive statements of possible futures which can beevaluated for accuracy in light of technological forecasting studies, development trendsand the political values or strategies which are intended to be met by using the technology.This component describes how the future will be, or what kind of technologies willemerge, based on ‘evidence’ available to the interlocutor.

2. Reaction—The respondent’s reaction to a particular use, e.g. BI wouldn’t feel comfort-able…I would feel like I was spying on her.^ These are local concerns and reactions basedon perceived uses, and cannot be disproven initially because only the respondent iscapable of saying how she will react (however, in due course these claims can provefalse if the respondent reacts differently than she expected once the technology isexperienced). Where a reaction is based on implausible futures or misunderstanding ofthe capacities of the technology, the respondent can be prompted to amend or clarify thereasoning behind their reaction.

3. Values—The moral beliefs or values which ground and explain the reaction, e.g. BI wouldfeel like I was spying on her…she was a very private person.^ Here the respondent isclaiming that ‘spying’ is wrong, in part because her mother is entitled to some degree of

6 Criteria were developed in this project to assess the success of particular instances of translation based on thecomprehensibility and acceptability of translated claims according to participants in the discourse. These criteriago beyond the scope of this paper, the key aspect of which is to suggest future avenues of conceptualdevelopment of discourse ethics as an account of the epistemic status of uncertain normative claims, and fordevelopment of a methodology of translation. For further details, see: Mittelstadt 2013 pp. 198-207.


privacy. This part of the statement expresses evaluations of right and wrong, alternativeconceptions of morality and ethical values. The movement from values or conceptions ofmorality to a reaction and normative claim can be critically questioned.

This model of the structure of uncertain normative claims reveals how translation can be putinto practice in analysing normative empirical data. Parts 1 and 2 are the source of uncertaintyin normative claims about the future—is the respondent’s description of the future plausible?Would interaction with the technology change how the respondent reacts to it (by changing hervalues)? Parts 2 and 3 constitute the normative truth content of an uncertain normative claim—so long as the description of the future in part 1 is credible, the reaction to it in part 2 based onthe values evident in part 3 must be taken seriously in discourse. In other words, assuming part1 is correct or at least plausible based on (for example) foresight research or systems currentlyin development, parts 2 and 3 constitute a normative claim with sufficient plausibility to betreated as hypothetically equivalent to other types of normative claims about the present. Thisis not to suggest the two have equivalent epistemic status, but rather that the uncertain claimcan be admitted to further discursive questioning with the confidence that it is a plausibledescription of the future based on research or future predictions with empirical credibility.

Some consistency can be seen between current and future technologies in terms of theirnormative dimensions. For instance, fear related to privacy violations resulting from inappro-priate sharing of data are equally applicable given current and future technological capacities.This consistency can thus be used as a basis to judge the future claim as legitimate; if the futureincident were to occur today, would the normative claim be justified? In this case, the temporalaspect of the incident becomes more or less irrelevant in terms of the credibility of the claim.While the future incident will undoubtedly have a unique character or degree of seriousness(e.g. much more sensitive data may be at risk in the future due to technological developments),and the user’s values or attitudes towards the technology may change over time from exposureto it (cf. Latour and Venn 2002), the consequences of the privacy violation can still beunderstood initially and expanded upon in understanding the normative importance of theclaim about the future.

For the three component parts to be open to discursive questioning, translation requiresinterpreting claims with due regard to frame of reference of participants in the discourse, so asto ensure the meaning of the concepts and terms employed in the translation are shared as faras possible (recognising that identical meaning is impossible). To be successful, the translationmust not only emphasise features deemed relevant or important by the translator, but explainthese features using a frame of reference (language, culture, history, concepts) that are familiarto the interpreter. As Habermas explains, the translation of religious beliefs in secular politicaldiscourse must discard controversial features such as Bfaith^ and Bmiracles^ which areunlikely to be considered relevant or important by members of society which do not sharecultural consistency with the person making the claim (Habermas 2011). The same can be saidfor uncertain claims made on the basis of foundational beliefs, rumours, misunderstanding ofthe capacity of emerging technologies, or perceived uses with little empirical support fromforesight studies, policy positions, or other indicators of future contexts.

Deconstructing claims in this way implies roles and responsibilities for participants indiscourses on the ethics of emerging technologies. ‘Facilitators’ of translation, or the ethicists,scientists, developers and others with a particular ‘expertise’ or knowledge base (that isrequired to mediate the discourses considering such technologies) have a responsibility toassist (non-technical) participants (e.g. citizens, policy-makers) in translating their uncertainnormative claims for acceptance into discourse. Facilitators wield considerable power in thisrole as their assistance may make the difference between the successful translation and thus


broader consideration of potentially controversial and at least uncertain claims. This mayimply a ‘duty of custodianship’ to pursue impartiality in supporting participants in translatingtheir claims. Conflicts of interest are therefore likely and problematic for facilitators assuminga dual role as participants in the discourse, meaning a separation of roles or the assistance ofthird parties as facilitators may be required to ensure translation is carried out fairly and inadherence with the ideals of a rational discourse.

As the deconstruction described here relies upon empirically grounded assessment of thetechnical and moral potential of the emerging technology being discussed, a duty is implied forfacilitators to keep up-to-date with the current status of development and proposed uses.Facilitators can thus make empirically grounded and transparent judgments to dismiss predicted‘characteristics of the future’ as unfeasible if, for example, the technology or motivations, socialor policy structures required for predicted uses both currently do not exist and are unlikely toexist in the future. A background in ethical theory and moral values and concepts which canground the claims made by participants in the discourse also appears necessary to understandand unpack the moral reactions and values upon which uncertain normative claims are built.

‘Participants’ likewise bear a responsibility to explain themselves and their opinions as faras possible, and a willingness to contribute as far as possible to translation of their uncertainclaims if they wish to participate in discourses about the future. Doing so aims to clarify claimsaccording to the criteria chosen in a particular discourse to admit uncertain claims—forinstance, it may be that claims lacking grounding in existing practices (e.g. current develop-ment, proposed uses of new technologies, ethical issues in related technologies or domains) arenot admitted to discourses about the future ethical acceptability of an emerging technology.7

These roles and responsibilities were enacted through discursive questioning of participantclaims and the background training of the researcher (‘facilitator’) in the aforementioneddoctoral research project, albeit without explicit explanation to the participant of their roleand the need for translation. Going forward in developing practical tools for future discourses,such questioning should presumably be included.

This account clarifies the components of an uncertain normative claim which are beingquestioned in discourse based on the TCA’s three validity claims. Questions over the validity ofeach part of an uncertain claim can raised, meaning uncertain claims can be treated the same asany other claim in discourse. Translation clarifies the epistemic status of uncertain normativeclaims by distinguishing the relative certainty and normativity of each part. Mutual understand-ing is facilitated by the location of common basic premises between the values grounding eachrespondent’s reaction to a potential future, even where these values lead to contradictorynormative claims. The problem of uncertainty is not eliminated in the sense that the respon-dent’s description of and reaction to the future may prove to be incorrect; however, the valuesgrounding this reaction, which speak to how the respondent would react to future scenarios, canbe understood and given consideration in public discourses which inform the governance ofemerging technologies. Further empirical research is undoubtedly required to validate andrefine such ‘acceptance’ and ‘dismissal’ requirements for assessing and admitting translatedclaims in discourse, and to refine the proposed role of discourse facilitators and participants.

7 Alternatively, it may be desirable to admit claims based on public fears or superstition to the discourse if theconcerns grounding the fears as seen as legitimate, for example by conceptualizing ‘fear’ as a form of mentalharm. The determination of appropriate criteria for translation relates to existing issues in risk management overthe weight given to such ‘empirically groundless’ fears about the future. While it may be possible to prescribe‘universal’ criteria based on Habermas’ Theory of Communicative Action, it may similarly be preferable todevelop criteria for accepting translated claims into discourse on a case-by-case or culture-by-culture basis. Thisproblem, which fundamentally addresses moral relativism, goes beyond the scope of the paper but is certainly anarea requiring consideration in future discourses.


5 Conclusions

Apart from the success or failure of the methodology of translated described above, the need toreconsider the epistemic status of uncertain normative claims used in ethical assessment ofemerging technologies and anticipatory decision-making remains clear. Further research isrequired to confirm discourse ethics as an appropriate foundation for an account of theepistemic status of such claims. The problem identified here requires further considerationbecause current techniques for the ethical evaluation of emerging technologies lack a clearaccount of the epistemic status of the evidence base upon which decisions are made.

An epistemological problem exists with all current attempts to base policy, governance,system design and other anticipatory action on normative empirical data which is gatheredunder conditions of uncertainty. The foundation of this problem lies in the relationship betweenfacts, norms and the future—if facts and norms cannot be derived from empirically observablephenomena and existing social practices (respectively), as is the case in considering the future,then they cannot be given the same epistemic status as facts and norms concerning existingphenomena. Despite this, uncertain normative claims should be taken seriously in craftingsocietal responses to emerging technologies, just as religious ideas should be taken seriously inpolitical discourse. In both cases the ideas must be translated to make their epistemic statusclear, and to reveal the evidence base upon which the claims rest. Normative claims about thefuture made must be taken seriously by policy makers, just as religious claims need to be givenattention in secular political discourses. An account of the epistemic status and treatment ofnormative empirical data in discourse has been suggested as a way forward to allow for suchclaims to be considered in public discourses through which their relative validity and influencemay be established.

To move forward, the methodological and validity requirements of ‘translation’ requirefurther development. While the necessary epistemological basis and concepts have beendescribed here, only initial suggestions have been made regarding the shape of a ‘methodologyof translation’. In particular, further development is required of tools and innovative methodswhich can assist participants in discourses concerning the ethical acceptability and restrictionor modification of emerging technologies with translation, including exploration of theresponsibilities and powers wielded by ‘facilitators’. Such practical advances serve a dualpurpose. Firstly, they may allow for ethical, legal and social norms to be embodied in technicalartefacts, which has been recognised as a necessity to give ‘propositional’ or conceptual levelexercises (as conducted here and in the ‘ethics of emerging technologies’ more generally)practical and noticeable impact on the technologies concerned (cf. Hildebrandt and de Vries2013). Secondly, they ensure translation itself adheres to the ideals of a rational discourse,meaning facilitators do not come to dominate the process. Such a power imbalance (andstrategic discourse) could occur if facilitators reject all claims with uncertain normativegrounding from the discourse by setting impossible standards for acceptance of translatedclaims into the discourse, or do not attempt to provide serious assistance to citizens intranslating (or unpacking) their uncertain normative claims. This issue mirrors Habermas’concern that even ordinary citizens will struggle to translate or find a secular basis formetaphysical claims (2011); however, the initial steps in ‘dissecting’ uncertain claims de-scribed above (see: Section 4.3.1) provide grounds for optimism. Establishing the potentialvalidity of the initial steps taken here and monitoring whether translation is possible in practicewithout succumbing to strategic discourses dominated by facilitators are both outstandingissues requiring further research and empirical validation going forward.

This work aimed to identify an often unspoken problem in current approaches to the ethicalassessment of emerging technologies, and related areas of future-oriented research which call


for anticipatory action at a social and political level. The paper described a conceptualapproach to establish the relative epistemic status of uncertain normative claims in discourse.In practice, the framework provides a way forward to clarify the validity of anticipatoryevidence-based policy and governance which relies upon uncertain normative empirical dataas evidence. The approach provides a way forward to evaluate the relative epistemic value ofuncertain normative claims, or their ‘validity’ in comparison to other types of claims consid-ered in discourse. By establishing the epistemic status of such claims the discourses throughwhich anticipatory action emerge can be defended as epistemologically sound, therebyproviding justification for the actions and decisions produced. Our account also suggests thatresearchers can base reasonable validity claims on normative empirical research on the ethicsof emerging technologies while uncertainty still prevails as long as such claims are translatedfor further discursive questioning. Anticipatory evidence-based policy-making and governanceshould therefore take heed of normative empirical data about the future, and take it seriouslyalongside other types of (observational) evidence in political discourse.

Acknowledgments The research leading to these results has received funding from the European Community’sSeventh Framework Programme (FP7/2007-2013) under grant agreements n° 230318 and 230602.

References

Ahmed A, Skogh G (2006) Choices at various levels of uncertainty: an experimental test of the restateddiversification theorem. J Risk Uncertain 33:183–196. doi:10.1007/s11166-006-0332-0

Baldwin C (2008) Family carers, ethics and dementia: an empirical study. In: Widdershoven G (ed) EmpiricalEthics in Psychiatry. Oxford University Press, New York, pp 107–122

Banks G (2009) Evidence-based policy making: what is it? How do we get it? Productivity CommissionCanberra

Birnbacher D (1999) Ethics and social science: which kind of co-operation? Ethical Theory Moral Pract 2:319–336. doi:10.1023/A:1009903815157

Black M (1964) The gap between Bis^ and Bshould^. Philos Rev 73:165. doi:10.2307/2183334Borry P, Schotsmans P, DierickxK (2004)What is the role of empirical research in bioethical reflection and decision-

making? An ethical analysis. Med Health Care Philos 7:41–53. doi:10.1023/B:MHEP.0000021844.57115.9dBrey P (2011) Anticipatory technology ethics for emerging IT. CEPE 2011 Crossing boundariesCagnin C, Keenan M, Johnston R et al (2008) Future-oriented technology analysis: strategic intelligence for an

innovative economy. SpringerCollingridge D (1980) The social control of technology. Macmillan, PalgraveCuhls K (2003) From forecasting to foresight processes—new participative foresight activities in Germany. J

Forecast 22:93–111. doi:10.1002/for.848De Vries M, Gordijn B (2009) Empirical ethics and its alleged meta-ethical fallacies. Bioethics 23:193–201. doi:

10.1111/j.1467-8519.2009.01710.xDepartment of Health (2011) Whole system demonstrator programme - headline findings - December 2011Dusek V (2006) Philosophy of technology: an introduction. Blackwell PublishingFloridi L (2010) The Cambridge handbook of information and computer ethics. Cambridge University PressFreestone D (1991) The precautionary principle. In: Churchill R, Freestone D (eds) Int. Law Glob. Clim. Change.

Graham and Trotman, London, p 447Friedman B, Kahn P, Borning A (2008) Value sensitive design and information systems. In: Himma K, Tavani H

(eds) Handb. Inf. Comput. Ethics. Wiley Blackwell, pp 69–102Gardiner SM (2006) A core precautionary principle. J Polit Philos 14:33–60. doi:10.1111/j.1467-9760.2006.00237.xGenus A, Coles A (2005) On constructive technology assessment and limitations on public participation in

technology assessment. Tech Anal Strat Manag 17:433–443. doi:10.1080/09537320500357251Georghiu L (2008) The handbook of technology foresight: concepts and practice. Edward Elgar PublishingGoodin RE (1983) Political theory and public policy. University of Chicago PressGrunwald A (2009) Technology assessment: concepts and methods. Handb. Philos. Sci. 9Habermas J (1981) Theorie des kommunikativen Handelns. Suhrkamp, FrankfurtHabermas J (1991) Erläuterungen zur Diskursethik. Suhrkamp, Frankfurt


http://dx.doi.org/10.1007/s11166-006-0332-0

http://dx.doi.org/10.1023/A:1009903815157

http://dx.doi.org/10.2307/2183334

http://dx.doi.org/10.1023/B:MHEP.0000021844.57115.9d

http://dx.doi.org/10.1002/for.848

http://dx.doi.org/10.1111/j.1467-8519.2009.01710.x

http://dx.doi.org/10.1111/j.1467-9760.2006.00237.x

http://dx.doi.org/10.1080/09537320500357251

Habermas J (1992) Moral consciousness and communicative action. The MIT PressHabermas J (1997) Between facts and norms: contributions to a discourse theory of law and democracy, new ed.

Polity PressHabermas J (1998) The inclusion of the other: studies in political theory. The MIT Press, CambridgeHabermas J (2008) Between naturalism and religion: philosophical essays. PolityHabermas J (2011) BThe political^: the rational meaning of a questionable inheritance of political theology.

Power Relig. Public Sphere. Columbia University Press, pp 15–33Heng MSH, De Moor A (2003) From Habermas’s communicative theory to practice on the internet. Inf Syst J 13:

331–352. doi:10.1046/j.1365-2575.2003.00144.xHildebrandt M, de Vries K (eds) (2013) Privacy, due process and the computational turn: the philosophy of law

meets the philosophy of technology (Hardback) - Routledge. RoutledgeHongladarom S (2005) Electronic surveillance in the workplace: a Buddhist perspective. Electron. Monit.

Workplace Controv. Solut. Idea Group Publishing, Hershey, PA, pp 208–225Hourdequin M (2007) Doing, allowing, and precaution. Environ Ethics 29:339Hughes J (2006) How not to criticize the precautionary principle. J Med Philos 31:447–464. doi:

10.1080/03605310600912642Hume D (2004) A treatise of human nature: being an attempt to introduce the experimental method of reasoning

into moral subjects. Penguin ClassicsJohn S (2010) In defence of bad science and irrational policies: an alternative account of the precautionary

principle. Ethical Theory Moral Pract 13:3–18Jonas H (1984) Das Prinzip Verantwortung: Versuch einer Ethik für die technologische Zivilisation. Suhrkamp,

FrankfurtJordan A, O’Riordan T (1999) The precautionary principle in contemporary environmental policy and politics.

Prot. Public Health Environ. Implement. Precautionary Princ. Island Press, Washington DC, pp 15–35Joss S, Bellucci S (2002) Participatory technology assessment: European perspectives. Centre for the Study of

Democracy, University of Westminster, LondonKant I (1827) Kritik der praktischen VernunftKant I (1870) Grundlegung zur Metaphysik der Sitten. L. HeimannKeynes JM (1921) Treatise on probability. Macmillan, LondonKlein HK, Huynh MQ (2004) The critical social theory of Jürgen Habermas and its implications for IS research.

Soc. Theory Philos. Inf. Syst. Wiley, Chichester, pp 157–237Knight FH (1921) Risk, uncertainty and profit. Houghton Miflin, BostonLatour B (1987) Science in action: how to follow scientists and engineers through society. Harvard

University pressLatour B, Venn C (2002) Morality and technology the end of the means. Theory Cult Soc 19:247–260. doi:10.

1177/026327602761899246Leget C, Borry P, De Vries M (2009) BNobody tosses a dwarf!^ The relation between the empirical and the

normative reexamined. Bioethics 23:226–235. doi:10.1111/j.1467-8519.2009.01711.xLyytinen K, Hirschheim R (1988) Information systems as rational discourse: an application of Habermas’s theory

of communicative action. Scand J Manag 4:19–30. doi:10.1016/0956-5221(88)90013-9Magnani L (2007) Morality in a technological world: knowledge as duty. Cambridge University PressMcMillan JH (2008) The possibility of empirical psychiatric ethics. In: Widdershoven G (ed) Empirical Ethics in

Psychiatry. Oxford University Press, New York, pp 9–22Mill JS (2002) Utilitarianism, 2nd revised. Hackett Publishing Co, IncMingers J, Walsham G (2010) Toward ethical information systems: the contribution of discourse ethics. MIS Q

34:833–854Mittelstadt B (2013) On the ethical implications of personal health monitoring: a conceptual framework for

emerging discourses. De Montfort UniversityMolewijk AC, Stiggelbout AM, Otten Wet al (2003) Implicit normativity in evidence-based medicine: a plea for

integrated empirical ethics research. Health Care Anal 11:69–92Moore GE (1993) Principia ethica: with the preface to the second edition and other papers. Cambridge University PressMorris J (2000) Rethinking risk and the precautionary principle. Butterworth-HeinemannMumford E (1996) Systems design: ethical tools for ethical change. MacmillanMumford E (2003) Redesigning human systems. Idea Group Inc (IGI)Musschenga AW (2005) Empirical ethics, context-sensitivity, and contextualism. J Med Philos 30:467–490Pantelopoulos A, Bourbakis NG (2010) A survey on wearable sensor-based systems for health monitoring and

prognosis. IEEE Trans Syst Man Cybern Part C Appl Rev 40:1–12. doi:10.1109/TSMCC.2009.2032660Popper KR (1959) The logic of scientific discovery. Basic Books, OxfordReuzel RPB, van der Wilt GJ, ten Have HAMJ, de Vries Robbé PF (2001) Interactive technology assessment and

wide reflective equilibrium. J Med Philos 26:245–261. doi:10.1076/jmep.26.3.245.3015


http://dx.doi.org/10.1046/j.1365-2575.2003.00144.x

http://dx.doi.org/10.1080/03605310600912642

http://dx.doi.org/10.1177/026327602761899246

http://dx.doi.org/10.1177/026327602761899246

http://dx.doi.org/10.1111/j.1467-8519.2009.01711.x

http://dx.doi.org/10.1016/0956-5221(88)90013-9

http://dx.doi.org/10.1109/TSMCC.2009.2032660

http://dx.doi.org/10.1076/jmep.26.3.245.3015

Roncaglia A (2009) Keynes and probability: an assessment. Eur J Hist Econ Thought 16:489–510Sandberg A (1985) Socio-technical design, trade union strategies and action research. In: Mumford E,

Hirschheim R, Fitzgerald G, Wood-Harper T (eds) Res. Methods Inf. Syst. IFIPColloquium Proc. NorthHolland, Amsterdam, pp 79–92

Searle J (1964) How to derive Bought^ from Bis.^. Philos Rev 73:43. doi:10.2307/2183201Sinn H-W (1980) A rehabilitation of the principle of insufficient reason. Q J Econ 94:493–506.

doi:10.2307/1884581Smith A (2007) The theory of moral sentiments - Adam Smith. Standard Publications, IncTran TA, Daim T (2008) A taxonomic review of methods and tools applied in technology assessment. Technol

Forecast Soc Chang 75:1396–1405. doi:10.1016/j.techfore.2008.04.004UNCED (1992) Rio Declaration on Environment and Development - United Nations Environment Programme

(UNEP). http://www.unep.org/Documents.multilingual/Default.asp?DocumentID=78&ArticleID=1163.Accessed 23 Nov 2012

Van Eijndhoven J, van Est R (2002) The choice of participative technology assessment methods. In: Joss S,Belucci S (eds) Particip. Technol. Assess. Eur. Perspect. Centre for the Study of Democracy, University ofWestminster, London, pp 209 – 234

Van Hooren RH, Van der Borne HW, Curfs LMG, Widdershoven GAM (2008) Providing good care in thecontext of restrictive measures: the case of prevention of obesity in youngsters with Prader-Willi syndrome.In: Widdershoven G (ed) Empirical Ethics in Psychiatry. Oxford University Press, New York, pp 153–171

Wagner WE (2003) TheBbad science^ fiction: reclaiming the debate over the role of science in public health andenvironmental regulation. Law Contemp Probl 66:63–133

Widdershoven G, Abma T, Molewijk AC (2009) Empirical ethics as dialogical practice. Bioethics 23:236–248Wingspread Statement (1998) The precautionary principle


http://dx.doi.org/10.2307/2183201

http://dx.doi.org/10.2307/1884581

http://dx.doi.org/10.1016/j.techfore.2008.04.004

http://www.unep.org/Documents.multilingual/Default.asp?DocumentID=78&ArticleID=1163

How to Shape a Better Future? Epistemic Difficulties for Ethical Assessment and Anticipatory...

Documents

Transcript of How to Shape a Better Future? Epistemic Difficulties for Ethical Assessment and Anticipatory...