Paper presented at the 4S & EASST Conference on Public Proofs science technology and democracy, Paris 25-28 August 2004 (31 augustus 2004)

The proof of the pudding is in the eating:
testing expertise in citizens panels and courts

1 Expertise an essentially contested concept
Expertise seems to have turned into an essentially contested concept (Marres and De Vries 2002: 184, Gallie 1956). Whoever claims to be an expert - or is consulted for her expertise - will encounter questions of validity and relevance. This, of course, is not a new development. However both the scale on which new expertise proliferates and the awareness of the necessity to draw on a plurality of perspectives and thus on a plurality of expertises, makes it impossible to revert back to monodisciplinary approaches claiming exclusive access to the solution of a specific problem. The traditional idea that problems of validity can be solved by checking on the constraints within the discipline or practice (Foucault 1971, Stengers) that has produced the expertise is upset by the interactions between disciplines that cause a permanent rebuilding of these internal constraints (f.i. the influence of epidemiology on molecular biology and genetics, Jasanoff 1995). At the same time the traditional idea that problems of relevance can be solved by listening to the experts is upset by the growing influence of laymen that are confronted by a multiplicity of expertise that produces contradictory claims on truth, reality and the solution of the problem at hand (Irwin 2001, speaking of the construction of the scientific citizen). In short, the traditional idea that laymen can defer their judgement on matters of fact to experts is turned around into the more challenging idea that matters of concern will be decided upon in a process of negotiation between those that are affected (Dewey’s publics, Dewey 1970) and those that claim relevant expertise. The turnabout from matters of fact to matters of concern (Latour 2004) of course implies that the concept of expertise is itself dependant on the matter at hand and will include those ’laymen’ that are expert on the basis of their experience of the matter at hand that concerns them since they (will) suffer/enjoy the consequences. Which raises questions about the difference between professional and non-professional expertise and between experiment and experience.

2 How to organize public proof: double constraints
A public proof can be understood as a public demonstration of the truth of a particular proposition on a matter of fact. In the academies and salons of the 16th - 18th centuries this type of public proof became the hallmark of a succesfull experiment and - in the course of time - of an objective truth. This type of public experiment involved both experts (those that designed and developed the artificial constraints that discriminate the experiment from experience) and laymen (even when these laymen were part of an elite, presumed to have better judgement - this was not a matter of mob rule, compare Latour’s Common knowledge). In the 19th century, after moving out of the salons and academies, into the universities (laboratories), this type of public demonstration changed into a more exclusive type of experiment. The public was restricted to experts that could testify/verify/falsify the process and the results on the basis of their exclusive access to specialized knowledge and training, without which the experiments would not make sense. Robert Merton’s community of scientists thus created two separate but interrelated problems: science is no longer a public undertaking, and science produces knowledge that is permanently under threat of contestation within the community of scientists. In a way scientific proof is no longer public proof. At the same time our lives depend on scientific facts, produced in the laboratory, proliferating in our common world (medicin, food, physical and virtual infrastructure etc.). Following Latour’s Politics of Nature, the question is how to bring the sciences that produce these facts back into democracy, how to organize public proof, not restricted to matters of fact (relevant in the laboratory) but to matters of concern (relevant in the field). (Compare Callon, Lascoumes and Barthes 2001).

In this paper the central question will be ’how to organize public proof’ in matters of concern, with special attention to the role of scientific expertise. Proof is taken to signify closure on the matter(s) at hand, and public is taken to signify both access to and participation in the construction of this proof by the relevant public. The relevant public is taken to signify all those that suffer and/or enjoy the consequences of the newly constructed world. Our point of departure will be that scientific facts have a provisional nature, due to their experimental origin: the black box that was fabricated can always be opened and put to a new test. This is not only the concern of scientists but also of those that will suffer and/or enjoy the impact of new scientific facts. Scientific facts - like any other factum - are stabilized constructions that have to be maintained, and will slowly or suddenly be transformed or replaced by other scientific facts. The introduction of new scientific facts into our common world is a matter of public concern and will demand another type of proof than the scientific proof of the laboratory, its inscriptions and its attachments within a specific scientific community. The black box that leaves the laboratory once it is taken for granted within the boundaries of one or more scientific disciplines, will have to be reconstructed into a black box that can survive in the field, while thus also reconstructing the field. The laboratory factum must generate loyalties, agreement and formal decision making in and of our common world. This often happens by attaching to commercial undertakings that were willing to invest in the experiments that created these new facts, such that these undertakings can gain a profit from their proliferation into the world. The stakeholders of this process seem to be limited to the researchers, those that seek a profit (enterprise, banks, insurance-companies) and those that must spend their resources to incorporate the new facts into their lives (consumers, government). In speaking of public proof the public is reintroduced as the main stakeholder: those that (will) suffer/enjoy the impact. The idea is that they should be heard or even have the last say in the matter (closure). The organisation of public proof thus involves a double set of constraints: those of robust scientific knowledge (however provisional) and those of the concerned public. These constraints are interrelated, precisely because expertise (and scientific knowledge) is an essentially contested concept.

3 Procedure: preuve and ?preuve
As lawyers we are tempted to answer the question of how to organize public proof in terms of procedure. Since we are also legal theorists this could open a black box on the meaning of procedure, but we shall restrain ourselves and use the term in a loose sense, such that it means little more than ’the way to organize closure’. In the history of legal procedure, especially of the criminal trial, two distinct ways to reach closure can be discriminated, best indicated by the French ?preuve (test, trial) and preuve (result of an inquiry, a proof).

?preuve is like a test, it involves a risk on both sides. Historically we are talking about the medieval mediative trial: a co-ordinate way to deal with uncertainties in a situation where there is no government, no monopoly on violence and thus no investigative power. As Foucault describes in his Brazilian conferences on ’la v?rit? et les formes juridiques’ (lecture III, Dits et ?crits 1970-1975, p. 570-588), ’in such a procedure there is no judge, no verdict, no truth, no investigation, nor any wittness to know who is speaking the truth. The decision, not on who has told the truth, but on who is right, is trusted to a contest, to the challenge, to the risk both are running’ (p. 570, translation mh). Another way of putting it is that we are in a jurisdiction that is voluntary, where proceedings are oral and where those that share jurisdiction form a society of peers. This means five things: first of all disputes and/or alleged violations of norms cannot be brought into court unless both ’parties’ wish to do so; second, the verdict of the court cannot be implemented by force upon a party; third, since the law is not written it is not conceived as something that pre-exists the trial and only needs mechanical implementation; fourth, the judge is a peer of the ’parties’, his ’authority’ depends on personal attributes not on a hierarchical relationship; fifth, as a consequence, the judge cannot find the law unless the parties agree about it, so the judge needs the co-operation of the parties (Glenn 2000: 56-86, Immink 1973, Salas 1992, Hildebrandt 2005).

Preuve is used here in the sense of the unilateral decision on matters of fact, made after an investigation. Historically we are speaking about the modern trial (the continental inquisitorial and the common law jury-trial that developed in the royal jurisdiction under the authority of absolute monarchies). It is a sub-ordinate way of dealing with transgressions of the royal legal order, and it presumes the royal monopoly on violence and the power and competence to investigate. The risk of the procedure is no longer with both parties; it rests entirely with the accused. Foucault (1975: 570) uses the term enqu?te to indicate the contradiction with the ?preuve. While the ?preuve speaks for itself: it is a test amongst peers that does not require a verdict to confirm its outcome (it is rather the consequence of a verdict in which the ?preuve is assigned to the parties), the preuve is a decision taken after an enqu?te is held. This is not a battle amongst peers. For Foucault the enqu?te is a legal procedure destined to find the truth, calling wittnesses to testify what they saw. This is - for him - the beginning of a rational procedure of investigation that holds the germ of the scientific revolution of the 17th century, in which humans are asked to testify what they have seen during the experiments in the salons and academies. The enqu?te is the legal way to find the truth. The procedure that uses the enqu?te or inquiry (the inquisitorial procedure) is no longer a means of solving a conflict between parties (like the mediative procedure that uses the ?preuve), but a means of establishing the truth. In common law jurisdiction the same shift can be seen at the end of the middle ages when the jurors are distinguished from witnesses, who are no longer allowed to serve as juror in the same case (Edmond and Mercer 1997). With Immink and Glenn I would like to shift the emphasis from truthfinding to the establisment of hierarchy. In a society without a state reaching consensus on matters of fact cannot be discriminated from reaching consensus on matters of concern (Latour Ein Ding ist ein Thing). The difference between a procedure that uses an ?preuve and one that uses a preuve is first of all that in case of a preuve a hierarchical relationship has developed between the parties and the judge, who is now representing a government. Those that share jurisdiction have become subjects to a king. Criminal jurisdiction is now compulsory and this means again five things: first, in criminal cases, the initiative to bring a case to court is with the government; second, the verdict is imposed upon the offender; third, especially in continental procedure, the - written, substantive - law precedes its implementation; fourth, the judge derives his authority from the government and stands in a hierarchical relation to the defendant/offender; fifth, as the judge does not depend on the co-operation of the defendant his verdict on matters of fact and law is unilateral, his judgement shares the authority of the state (Immink 1973, Glenn 2000, Salas 1992, Hildebrandt 2005).

Taking inspiration from L?vy-Bruhl (without endorsing his analysis of the irrationality of the archa?c proof (the ?preuve)) we could understand both preuve and ?preuve as mechanisms to establish closure in case of uncertainty (L?vy-Bruhl 1964: 15). The beauty of this description is that it highlights the uncertainty. Public proof is a thing called for in matters of uncertainty. ?preuve solves the uncertainty by organizing a - ritualised - (con)test, preuve solves the problem by organizing an investigation. ?preuve stands for co-ordinate justice, preuve for subordinate justice. ?preuve is premodern in the sense that a distinction between what is true and who is right cannot be made and is of no interest. Preuve is modern in the sense that it objectifies, puts matters in a vertical relation: the enqu?te presumes a subject (originally the king) that investigates an object (his subject).

4 Criminal procedure in the democratic constitutional state
After Foucault has described, the ’advance’ of ?preuve to enqu?te, from ritualised contest to judicial truth-finding, identifying the judicial enqu?te as the germ for scientific experimentation (science in action), he moves on to a third type of procedure: the examen, which is particular for disciplinary society and exemplary for the methodology of the social sciences as they developed in the 19th and 20th centuries. He thereby puts a finger on the co-existence of disciplinary practices and judicial procedure, contrasting the logic of control of the one to the rethoric of justice of the other. Elaborating on this he highlights the contrast between the emergence of the prison that fits so well in the logic of a disciplinary society, while it seems incongruent with the rethoric of classical legal theory.

Other than Foucault, though without wishing to invalidate his analysis, I will follow judicial procedure from the mediative trial of the peers in a society without a state, to the inquisitorial trial of the subjects of an absolute monarch, to the ’fair trial’ of citizens in a democratic constitutional state. Though I would agree that the logic of disciplinary practices can be at odds with the logic of the ’fair trial’, this makes it all the more important to examine the logic of this ’fair trial’ because it can restrict the unlimited exercise of control (whether by state or other agencies). Precisely because of the totalitarian logic of disciplinary practice the ’fair trial’ cannot be taken for granted.

In the ’fair trial’ subordinate and co-ordinate justice are integrated: while the defendant is object of investigation, she is also a legal subject, compensated for the asymmetries inherent in the criminal trial by an artificial ’equality of arms’ that will enable her to contest the charges. Though the relationship between judge and defendant remains vertical, enabling an investigation (enqu?te) and a unilateral verdict that shares the authority of the state, the relationship between defendant and prosecution is both vertical (due to investigative competence of the prosecution) and horizontal, enabling a contest on the matters of fact and norm. Contradictory procedure - both in its adversarial and in its continental form - integrates contest and investigation, or preuve and ?preuve, thus preparing closure on a matter of concern for all those that share jurisdiction.

The process-model of the ’fair trial’ demonstrates a way of organizing public proof: the point of the criminal trial is to reach closure in a matter of uncertainty that concerns all that share jurisdiction (since the legal certainty that is established has legal consequence for all that share jurisdiction); defense and prosecution are given the (legal) means to participate in the construction of this closure, whereby the prosecution represents all that share jurisdiction (the relevant public); the public-ness of trial and verdict assures access for those that are affected by the closure.

4.1 ’Fair trial’: laymen and experts in court
The artificial - legal - construction of closure in the ’fair trial’ confronts laymen with legal experts: defendant, witnesses and jury are laymen as far as professional legal expertise is concerned. In the continental trial closure is performed by the judge as legal expert, while in the adversarial trial the jury may (co-)perform closure in matters of fact and/or sentence. Apart from this interesting participation of laymen and legal experts the ’fair trial’ often involves other experts that are invited to contribute to the establishment of matters of fact. As Sheila Jasanoff (1995) expounds in her Science at the bar expert-evidence ’in court’ not only affects the case at hand and the legal matter to be decided: the contradictory setting in which expertise is tested often has a profound impact on the course of science itself. She argues ’that the cultures of law and science are in fact mutually constitutive in ways that have previously escaped systematic analysis’ and professes that both institutions ’jointly produce our social and scientific knowledge, and our relationships with technological objects’ (Jasanoff 1995, p. 8). In a court of law the expert not only contributes to the public proof of factual guilt, but the procedure itself challenges the authority of scientific expertise - case by case - and often ends up constructing new constraints within a specific field of expertise (compare Edmond and Mercer 1997).

Following the distinction between ?preuve and preuve one could say that expertise typically belongs to a trial in which investigation (enqu?te) is the main instrument to attain closure, while in a trial that is exemplified by the ?preuve scientific expertise does not seem to fit. The mediative trial seeks means to ritualise a contest or to pacify the parties, while the inquisitorial or even adversarial trial seeks to establish the transgression of a legal rule in order to sanction it. While the mediative trial is not concerned with matters of only fact (it does not practice the distinction between fact and norm), the inquisitorial and adversarial trial are based on the assumption that the historical truth can be established (in the case of an offence) and that this truth will legitimize the exercise of the ius puniendi. Expertise is part of the investigation (enqu?te) aimed at discovering this historical truth. Two points have to be made here. First of all, the adversarial trial seems to have already integrated a degree of co-ordinate justice into its structure. Even when in the end the adversarial anglo-american trial performs closure by a unilateral decision that shares the authority of the state, and therefore cannot be termed a mediative trial, it does incorporate elements of a contest. The common law trial is historically less focused on implementation of a pre-existing substantive law than on the resolution of conflict, less focused on legal rules than on the case at hand. It should be no surprise that the ’fair trial’ is indebted to a certain extent to the adversarial trial, as the ’fair trial’ also integrates vertical and horizontal elements with the purpose of creating some equality of arms. Though the contradictory nature of the ’fair trial’ is not equivalent with the adversarial structure (that is built to accommodate the jury), as we have seen above it does create room for a contest on matters of fact and norm.

The position of an expert can now be viewed from the perspective of both ?preuve (contest) and preuve (investigation). As far as some kind of contest is integrated in the ’fair trial’ experts can expect contestation in trial: testifying on the basis of one’s expertise then implies taking the risk of being refuted. As far as the trial is entirely based on the logic of authoritative investigation the - court-appointed - expert will be consulted by the judge, who will defer to the expert opinion in his judgment, presuming the expert has access to an objective truth. Thus ?preuve implies possible contestation of expertise, while preuve tends to imply consultation and deference.

4.2 Examples: use of statistical evidence in court
To give an example of the role of expert evidence in court we shall briefly look into some cases in which statistical arguments are presented as evidence in a criminal case, regarding the question whether certain incidents can be attributed to chance. The object of this exercise is not to analyse how this evidence is (or is not) used to construct legal proof, but to find out to what extent this expertise is tested and/or consulted and in which way it could contribute to the construction of public proof. We take it that legal proof is an example of public proof, though public proof refers to much more than just legal proof.

In the Netherlands two cases were recently brought before the courts (Court of Appeal of Amsterdam 7th October 2003, LJN nr. AL7736; Court of Appeal of The Hague 8th June 2004, LJN nr. AP 2846) in which a defendant was charged with murder because death and/or life-threatening incidents had occurred on a number of occasions while the defendant was on duty (as a nurse in a hospital; as a caretaker in a daycare for children). The suspicion that these incidents may have been caused by the defendant was raised by the surprising amount of cases happening during the duty of one person, without similar events taking place during the duty of other staff-members. In both cases the court consulted a statistical expert with the question whether this amount of incidents could be a matter of chance. Perhaps this is a very strange question, because a categorical answer seems out of the question. The only thing a statistician can do is to calculate, on the basis of certain presumptions, what is the chance/probability that a nurse or caretaker is confronted which such an unexpected high number of incidents. As was to be expected the probability was calculated to be extremely low but the interesting question the court had to then answer was what should be the meaning of this low probability for the construction of legal proof. It should be obvious that a person cannot be convicted for the murder of one or more patients on the basis of only these kinds of statistics. As the courts decided, legal proof in such matters must be based primarily on evidence that the defendant has caused the incidents (medical expertise, eye-witnesses etc.). While the court in Amsterdam held that the statistical argument could count as additional evidence (leading to legal proof of murder or attempted murder if other evidence was available), the court in The Hague did not make use of the statistical argument at all. By the time this last court decided the cases had attracted wide media coverage, including a far-reaching discussion amongst statisticians about the validity and relevance of their expertise in such cases and especially about the methodologies to be used. The court of appeal in The Hague invited both the consulted expert and other statistical experts to clarify their position in court. A battle ensued between Bayesian and non-Bayesian statistics, and its relevance for the question at hand: can the correlation between the defendant’s duty and the incidents be due to chance? The discussion evolved not only in newspapers or newsmagazines but also in the professional literature of both legal and statistical professions (Elffers 2004, Hildebrandt 2004, Dirk van Delft 2004, Stator 2004).
Several points are at stake: is the statistical method used adequate in these kind of cases, is it applied correctly and what is the relevance of the result for the construction of legal proof? In fact it turns out that these questions touch on further issues. Bayesian and classical statistics are not merely a matter of methodology, they touch on epistemological and ontological questions that do not only concern the outcome of the calculations but also the relevance of the outcome for the matter at hand. If most important things in life turn out to be highly improbable but nevertheless do occur, one could even ask the question if any relevance should be accorded to these kind of calculations. In a similar case (Sally Clark, see www.sallyclark.org.uk for a link to the judgement) the chairman of the Royal Statistical Society wrote that in cases such as this the relative likelihood should be calculated of alternative hypothesis, that are both highly improbable: f.i. the likelihood that the incidents are due to chance and the likelihood that they are due to murder (both calculations presuming there is no causal connection between the defendant and the incidents, any other presumption would be begging the question). The absurdity of this comparison indicates that the common sense notion that such incidents cannot be due to chance can be confirmed by complex calculations but that this leaves us with no more certainty than before. The need to calculate the probability of alternative hypothesis can of course be a good way to relativize our common sense, but as long a the judge realizes that a conviction on the sole basis of suspect co-incidence is not possible, we do not seem to need a statistician to give an expert opinion. For the construction of legal proof we need another type of evidence.

What is interesting about these cases is the relation between the construction of legal proof in court and discussion it provokes about the role of statistics in the construction of public proof. By being faced with a concrete problem asking for a concrete solution, the presumptions and implications of the use of statistics are exposed to a wider public and the controversial nature of conclusions drawn on the basis of statistical calculations refutes easy refuge into statistics as conclusive evidence. Though one cannot and should not extrapolate directly from a specific criminal case to general statements about the role of statistics in society, the cases are important in sensitizing larger publics to the complexities of (statistical) expertise and its use in the construction of public proof.

5 Participatory Technology Assessment (pTA)
The relationship between science and public policy (public used in the sense of governmental) is complex and - like the relationship between science and law - in many ways mutually constitutive. Both the natural and the social sciences seem to provide a lot of expert advise based on statistical calculations. As we have seen above, thinking along foucaldian lines, the relationship between the emergence of statistics as one of the most pervasive methods of the social sciences and the emergence of surveillance society go hand in hand with the advance of disciplinary practices. These practices co-exist - as we have seen above - with the judicial practices of the rule of law, exemplified by the logic of the ’fair trial’. Within the field of pTa some scholars hold that pTA actually offers an interesting alternative for the traditional practices of social research aimed at informing public policy. These traditional practices, like opinion polls and focus groups, depend on statistical arguments that are interwoven with the rethoric of representation. They are used in research for public policy but are also part of the tool-box of any marketing researcher. One could say that they perform a disciplinary function in a very much refined and highly sophisticated manner. Their logic is no longer based on normalizing people into preconceived categories, but on a continuous re-assessment of categories, allowing people to move around from one categorie to another but stil and continuously monitoring their attachments and thus retaining an intelligent control over what goes on in the world. The panoptic nature of this logic lies in the fact that this monitoring is input for new market-strategies or new public policy, while the citizen that is monitored is not invited to participate in the construction of these strategies or policies; she is not put in a position to monitor and impact the construction of public policy or commercial strategy.

Hereunder we will contend that the presumptions about representativeness inherent in both policy- and marketing research are problematic and that the construction of public proof in the fields of both public policy and commercial undertaking can be refined and advanced by integrating pTA practices. We wil however argue that pTA can rather easily be used to legitimize policy-strategies, because - other than the ’fair trial’ it has no formal structure embodying the checks and balances of the democratic constitutional state. To clarify this point we will compare the way both a legal jury and a pTA citizens jury contribute to the construction of public proof.

5.1 Representing public opinion/consumer preference: survey and/or pTA
Both public policy-makers and commercial undertakings want to be informed about the opinions/preferences of their target groups. One - traditional - way of getting this information is a kind of market research in the form of opinion polls or focus-groups. The issues are formulated by the researcher and the answers formatted in a quantifyable way. This means that they can be aggregated into percentages of citizens/consumers who agree or disagree with a set of opinions/preferences. The people that are thus researched are the object of research and their opinions and preferences are represented by the researcher on the basis of these aggregated statements. In fact, as the House of Lords Select Committee on Science and Technology is quoted as saying in relation to a report on the Public Consultation on the Biosciences: ’despite its name, we see the exercise as closer to market research than to public consultation’ (Irwin 2001:13). Interestingly the consultation was performed by one of Britains best known market-research companies MORI (Market & Opinion Research International), combining qualitative discussion groups with a larger quantitative survey. The exercise, as described by Irwin, demonstrates how qualitative research that takes its research object serious as a subject - as a citizen - can be integrated with quantitative research into a large-scale project that seems to satisfy both the ends of statistical representativeness and the need for more participatory forms of research. The opposition is then not only between qualitative and quantitative research, but between research that registers opinion/preference and research that facilitates learning and particpation. This integration however also runs the risk of adding legitimatization to policy-decisions that are nevertheless again based on ’hard numbers’, since policy-makers can interpret the pTA research in terms of ’subjective’ perceptions that must be corrected by better information on scienctific advance.

In a short but very informative article Tom Wakeford (2002) contrasts three types of representation: statistics (in the social sciences), voting (in politics) and common sense and/or informed consent (in juries). In relation to the last type of representation he writes about the citizens jury: ’Like a legal jury, the cornerstone of a citizens jury is the belief that once a small sample of a population has heard the evidence, their subsequent deliberations can fairly represent the conscience and intelligence of the community. This age-old reasoning contrasts with today’s most common quantitative and qualitative methods for representing the public’s view: the opinion poll and the focus group.’ Referring to the Magna Carta of 1215 he confirms the right to a trial by peers as a specific interpretation of democracy, that takes representation serious on another level than aggregated - static, decontextualized and ahistorical - votes or consumer choices. Interesting differences between voting, consumer choice and pTA are the structured learning process that is part of pTA and the fact that it aims at construction of informed consent. Rather than just registering an opinion pTa facilitates the construction of opinion, and rather than just registering individual opinion pTA facilitates the (re)construction of informed common sense (compare Wynne 2001:470). We do not imply that registration of individual opinion embodies nothing but a disciplinary practise, nor do we deny that pTA can be used to as windowdressing and/or symbolic use for the legitimisation of preconceived policy (Glasner 2001, p. 45). However, both can be put to good use, if democracy is understood to be more than majority rule and if checks and balances like those of the rule of law are integrated.

5.2 Legal jury and pTA jury: comparison of constraints
We shall now compare the legal jury in ’fair trial’ to the citizens jury in pTA (Wakeford 2002) on three points: first of all we will consider the framing of the issue, second we will look into the burden of proof (presumption of innocence and principle of precaution) and third we will look into the way the jury gains access to information and puts expertise to the test. The comparison is in many ways equally relevant for continental procedure, but the involvement of a panel of laymen in both adversarial procedure and citizen’s jury pTA makes it easier to find the relevant differences since the similarity is obvious. Regarding the testing of expertise in court, the judge is as much a laymen as the jury - in that way the comparison is instructive also for continental procedure.

5.2.1 Framing the issue
To begin with, the legal jury is faced with a closed question: guilty or not guilty. This seemingly simple question involves a whole range of legal and factual determinations that are intertwined (f.i. guilty of what?) and in the course of the process expertise may be hired to elucidate on matters of fact, all of which is aimed at answering this straight forward question with a simple ’guilty’ or ’not guilty’. In the mean time it should be understood that in criminal cases the presumption of innocence determines that as long as the question cannot be answered with a hight degree of certainty, the verdict will have to be ’not guilty’. Regarding the question to be answered it is also important to note who has the competence to formulate the question: in both common law and continental law the prosecutor has a monopoly on the articulation of the charges, while he is constrained by the legislator (or the common law) as to what charges he can possible file. The pTA citizen jury may or may not be faced with a question they cannot change and the degree to which the question is open to a plurality of answers and/or an elaborate answer depends on the setting: possibly the answer will be formulated in a further set of questions, to be answered before a certain technology should be implemented. This openness is at once an advantage and a disadvantage. It means that in some cases citizens can participate in the articulation of the question, while in other cases the question can be formulated in a way that excludes discussion of preliminary questions. Glasner (2001) for instance gives the example of a jury that addresses the following question: ’What conditions should be fulfilled before genetic testing for people susceptible to common diseases becomes available on the National Health Service (NHS)?’ Obviously the question whether genetic testing in itself is a good thing is passed over. Interestingly the commissioning body was a large transnational pharmaceutical corporation. But even if a citizen’s jury has room to reformulate the question or take on other questions it deems relevant, this openness may be a disadvantage because it may lessen the chance to have an impact on policy-makers. As Pimbert and Wakeford note, pTA procedures created ’from below’ (civil society instead of government) often take more space to frame the questions, while as the same time they have ’relatively weak links with the formal policy process’ (Pimbert and Wakeford 2001:26).

5.2.1 Burder of proof
A second point of consideration is that in the criminal trial a less than high degree of certainty does not count as legal proof, due to the presumption of innocence; in matters of policy the principle of precaution could fulfill a similar role, meaning that as long as citizens are not convinced of the desirability and/or necessity of certain technologies while expecting serious problems, these technologies should not be introduced (Wynn 201:466). However the principle of precaution is not usually applied in relation to the kind of ’proof’ constructed in a citizen’s jury, but rather in relation to a traditional risk analysis that is based on technological and scientific expertise (Wynne 2001). While the legal jury has the competence to create legal consequence in the matter of guilty/not guilty, the citizen’s jury has no such competence and depends on the agenda of those it advises.

5.2.3 Putting expertise to the test
A third point to consider when comparing legal and pTA citizen juries concerns the information they access during trial or delibiration. A legal jury has no opportunity to actively intervene in the questioning of witnesses or expert-witnesses, they also lack the competence to invite experts. For all this they depend on the parties involved. Their task is to listen and then deliberate amongst themselves untill a decision is reached. During their service as jurors they are not allow to speak to witnesses or public, they are segregated and forced to reach consensus within a limited amount of time (sometimes juries vote, and sometimes a jury ’hangs’ meaning they cannot reach a unanimous or majority verdict). The citizen jury often decide on hearing specific experts and can take the lead in questioning them. This may seems advantagous, but this will depend to a large extend on the composition of the jury. While in court the parties are clearly stakeholders with opposite interests, trying to get to the bottom of things, a jury that does not incorporate the greatest diversity of relevant stakeholders may not go to the bottom of things, because certain aspects of the issue are not brought forward with enough force. With Pimbert and Wakeford we could say that in matters of public concern ’the concept of stakeholder should include those who are ’stake-less’, having been marginalised by prevailing socio-economic forces’ (2001:27). If - again - we understand a public in the sense of Dewey as those who suffer/enjoy the consequences of a certain policy, than this public (whether or not marginalised) should be present in the jury to remind everyone what is ’at stake’. The ’fair trial’ is organised in such a way that expertise can be put to test if this is relevant for the outcome of the case. The contradictory nature of the trial is ensured by legal means - with legal consequence - and the involvement of legal professionals on both sides should guarantee that these instruments of contestation are not lightly waived. The opposing interests of both parties thus form an intelligent network of checks and balances, that should prevent for instance the state from exerting undue influence on the defendant to plead guilty. The fact that in real life the pressure to enter plea negotiations has risen may be an indication that the relation between disciplinary practice and ’fair trial’ is out of balance. In the case of the citizens jury many different configurations are practised, depending on who is funding, the budget, what is at stake, which initial information is made available, who decides on the invitation of experts etc. Pimbert and Wakeford differentiate between pTA created from above (convened by government agencies) and those created from below (convened by civil society organisations), whereas both the framing of the questions and the contradictory nature of the issues may differ. We would like to contend that other than both pTA set up from below and pTA set up by commercial enterprise, a democratic constitutional government is constrained to invest in a sustainable structure that incorporates the checks and balances needed to reach an interesting and legitimate outcome. In that sense pTA - just like the ’fair trial’ - needs a strong state, one that lends its authority to contestation of its own authority (paradox of the constitutional democratic state).

6 Closure: public proof in court and pTA?
According to a standard story about the relationship between science and law, Science is characterized by a quest for certainty while Law is sacred, decisional and authoritative. Building on the inherently controversal nature of scientific knowledge we propose a different approach: while science can only deliver provisional knowledge, law performs closure. Legal procedure incorporates the fact that it has to perform closure under conditions of uncertainty; it organises public proof in a way that at once legitimizes and lends authority to decisions with legal consequence. In the ’fair trial’ this legitimization is based on both adherance to democratic legislation and the contradictory structure of the trial. On the one hand, if violation of a legal rule is established (preuve) punishment or compensation is awarded; on the other hand the establishment is prepared by a contest (?preuve) between prosecution and defendant that are given the means to have a ’fair’ fight. The public nature of the ’fair trial’ consists in the fact that all that share jurisdiction (and thus suffer/enjoy the consequences of a court’s decision) have access to the construction of the proof, while those that suffer or enjoy direct impact of the verdict participate in the construction of the evidence.

Participatory technology assessment (pTA), or deliberative and inclusionary processes (DIPs) as some authors call them (Pimbert and Wakeford 2001), aims to contribute to the formation of public knowledges regarding emerging technologies (compare TAMI:4). In many ways the assessment includes and advances public understanding of science (Wynne 1995), meaning the reconstitution of scientific knowledge in contexts outside the laboratory. According to the final report of the European project on method and impact of technology assessment (TAMI) ’TA should be seen as a necessary link in the process of decisionmaking and not as a subsitute for it’ (TAMI:12). In legal procedural terms pTA is focused on the construction of evidence, leaving the decision on proof to the political process. Other than the proceedings of a ’fair trial’ pTA prepares evidence for policy and legislation, not for a judicial verdict. The outcomes of pTA practices form just a small part of the advise given to policy-makers and legislators, who are by no means obligated to base their decisions on the findings of pTA. Summing up we could say that while legal procedure specializes in closure, pTA seeks to influence those that perform closure in the public sphere.

As to the publicness of ’fair trial’ and pTA, in the sense of transparency, access and participation, both work with a limited amount of actors that actually participate in the process. In both cases these actors do not represent the ’general public’ in the traditional meaning of parlementary democracy. It is obvious that this traditional (modern, individualist, rationalist) conception of representation has become controversal and that pTA is one of the attempts to develop new democratic practices that can transform democracy into a mixture of classical representation, rational deliberation and inclusionary participation. If we take a public to be formed by those that are affected by a certain policy, decision and/or technology, pTA does not necessarily choose its participants from the different publics or stakeholders that (will) suffer or enjoy the consequences (including ’stake-less’ stakeholders, as Pimbert and Wakeford 2001:27 suggest). A citizen jury or panel may very well be composed of citizens that form a cross section of society without being directly or even indirectly ’interested’ in the matter at hand. Often citizens are deliberately chosen that have no prior knowledge or interest in the subject-matter, so that they can start on equal footing though from very different perspectives. In that case it seems that pTA is looking for a reconstruction of the common sense or the public interest, in a way that fits the legitimization of the trial jury. Edmond and Mercer (1997:331) quote Hans: ’The jury’s primary function is to infuse community values into the legal system by interpreting legal standards and specific factual patterns according to changing community norms of conduct and justice’. A jury can thus be understood as a public under construction, while reconstructing the common sense on certain legal issues. In the same way pTA citizen participation can be seen as an instance of public reconstruction: meaning both the construction of a specific public and the reconstruction of the common sense.

Michel Callon, Pierre Lascoumes, Yannick Barthe, Agir dans un monde incertain. Essai sur la d?mocratie technique, Paris: Seuil 2001

Dirk van Delft, Een schijn van kans, NRC 13 maart 2004, bijlage Wetenschap & Onderwijs, p. 41

John Dewey, The public and its problems, Chicago: The Swallow Press (1927) 1970

Gary Edmond and David Mercer, Scientific Literacy and the jury: reconsidering jury ’competence’, Public Understanding of Science (6) 1997, p. 329-357

Henk Elffers, Bij toeval veroordeeld’?, NJB 2003, p. 1812-1814

Michel Foucault, ’La v?rit? et les formes juridiques’, idem, Dits et ?crits (II) 1970-1975, Edition Daniel Defert, Fran?ois Ewald, avec la collaboration de Jacques Lagrange, p.538-648

Peter Glasner, Rights or rituals? Why juries can do more harm than good, PLA Notes (40) 2001, February, p. 43-46

M. Hildebrandt, ’Trial and ’Fair Trial’: from Peers to Subjects to Citizens’, in: Antony Duff, Sandrea Marshall, Victor Tadros, Lindsay Farmer (ed.), The Trial on Trial II. Judgement and Calling to Account, London: Hart 2005

M. Hildebrandt, Wetenschap in rechte, TREMA 2004 april Special ’Deskundigen in het rechterlijk proces’, p. 187-196

Alan Irwin, Constructing the scientific citizen: science and democracy in the biosciences, Public Understanding of Science, (1) 2001, p. 1-18

Sheila Jasanoff, Science at the Bar. Law Science and Technology in America, Cambridge Massachusetts, London: Harvard University Press 1995

Sheila Jasanoff, The Life Sciences and the Rule of Law, Journal of Molecular Biology 2002 (319), p. 891-899.

Simon Joss, Arthur Brownlea, Considering the concept of procedural justice for public policy- and decision-making in science and technology, Science and Public Policy (26) 1999-5, p. 321-330

Henri L?vy-Bruhl, La preuve judiciaire. Etude de sociologie juridique, Paris: Marcel Rivi?re 1964

Stator, periodiek van VVS, themanummer ’Statistiek in de rechtszaal’, juni 2004

TAMI: Technology Assessment in Europe: between Method and Impact, final report 2004, te downloaden bij http://www.ta-swiss.ch, onder projecten (overige TA projecten)

Tom Wakeford, Citizens Juries: a radical alternative for social research, Social Research Update, (37) summer 2002

Brian Wynne, Creating Public Alienation: Expert Cultures of Risk and Ethics on GMO, Science as Culture, (10) 2001-4, p. 445-481

Brian Wynne, Public understanding of science, S. Jasanoff, G.E. Markle, J.C. Petersen and T. Pinch (eds.), Handbook of Science and Technology Studies, Thousand Oaks, CA: Sage 1995