Paper presented at our Colloquium Testing Expertise on 21st October 2004
The discussion at the end of the day on the significance of mathematical proof outside the domain of mathematics reminded me why I took the non-state society as default position to look into the transformation of proof as a means to reach closure. I have slightly revised the text to make this more explicit.
Provisional version, not for quotation. Do only refer to the final published version: Hildebrandt, M.,The Trial of the Expert: Èpreuve and Preuve, The New Criminal Law Review 1 2006 (1 )

he trial of the expert: ?preuve and preuve

This paper works out the historical difference between testing a case (risky business, presuming uncertainty) and proving a case (claiming objective knowledge, presuming certainty is possible). This historical perspective will then be used to highlight the different attitudes to expert knowledge in court: deferring to or trying of science in court? Special attention will be given to the establishment of legal certainty in relation to scientific uncertainty.

Is the expert on trial? And/or, is expertise on trial? Today? Yes, it seems to be the case. Not only do experts from different disciplines confront those that asked their advice with incompatible conclusions, not only do experts within the same discipline disagree about diagnosis and solutions of problems put before them, the question itself who is an expert with regard to a particular problem can no longer be solved by asking the experts. There are too many of them, they might not agree, and possibly the expertise needed for your problem still has to be invented. Besides, after hiring the expert, how can you decide whether her advice is relevant and/or valid. Is it not the case that if you could answer that question you wouldn’t need an expert in the first place? [I don’t think so, but the question does pop up]

In court, these questions are not only of theoretical interest, but an everyday affair that cannot wait for a definite answer. So judges, juries, prosecutors and council for the defense have found ways to deal with these problems as they occur. Creating the path as they walk it, crossing their bridges when they come to them. This is very interesting, because it turns out that courts - forced by the deadlines that provide our societies with legal certainty (or closure) - find ways to question the experts up to a point where a legitimate common sense is constructed around the problem at hand. This sensus communis is not constructed by the experts but emerges during the whole process of testing the expertise in the light of the case that has to be decided, in the interactions between the actors that people the court. However, the judge - or jury - decides, the deadlines mentioned stipulate that we cannot wait for spontaneous consensus to occur, so the judge - or jury - performs this thing for us: this thing being the establisment of facts (intertwined as they are with legal norms).

It is very interesting that the term fact seems to have originated in a legal context. In the continental legal tradition the factum was the formal document representing what law presumed to be true (for the limited purposes of law) (Glenn 2004:49). The awareness that the historical truth about a specific human action will often remain debatable and thus needs formal establishment to be treated as if true, is far removed from the conflation of facts, truth and reality as so often happens in the context of a modernist belief in objective reality. In the common law we find an even more interesting history of the term fact when we look at the phrase ’matter of fact’. According to Shapiro (2000:9-11) this phrase did not refer to establish facts but - quite on the contrary - to the issue as placed before the jury. It referred to an uncertainty, still to be decided upon. It is only at the end of the 17th century that in the discourse of the natural sciences the term fact begins to refer to an already verified truth: ’Indeed, one of the great changes that occurred over the course of two centuries in some cultural arenas was the transformation of ’fact’ from something that had to be proved by appropriate evidence to be considered worthy of belief to something for which appropriate verification had already taken place’ (Shapiro 2000:31).

In this contribution I will try to set a stage on which the different ways of establishing fact and of providing proof can be further explored. I will do this by looking into the difference between ?preuve (contest) and preuve (proof) from a legal perspective and I will claim that historically speaking proof is not a mathematical concept but a legal precept, concerning practical wisdom rather than theoretical knowledge. Proof and fact originally refer to matters of uncertainty, that seek closure in the form of legal certainty. The way this closure is prepared depends on the way society is organised. Following Foucault’s Brazilian lectures on ’The truth and the legal forms’ of 1973 I will discriminate between (1) a society without a state, in which the ?preuve functions as a ritualised contest between peers; (2) the absolute state in continental Europe of the 16th/17th centuries, in which the enqu?te functions in an the inquisitorial trial that objectifies the defendant as subject of the absolute king and lastly (3) the formal ’Rechtsstaat’ of the 19th century in which the law seems to function as a means to control its citizens, while the exam is used as a tool to measure and categorize citizens. For Foucault this exam is closely embedded in what he calls disciplinary practices, that measure deviance from a norm while producing it (normalisation). According to Foucault these practices form a kind of ’contre-droit’ that erodes the potency of Enlightenment liberalism and legalism. I will argue that Foucault’s opposition of strict legalism to disciplinary practices is interesting, but not the end of the story. On the basis of a relational conception of law (in connection with a substantial understanding of the ’Rechtsstaat’) I will argue that the fair trial is both more and less than a hallmark of legality and liberalism and that it facilitates the testing of evidence (expertise) on a case to case basis, thus counterbalancing and disorganizing the possibly totalitarian effects of diciplinary practices. In that sense this ’fair trial’ is an interesting example of how to deal with expertise from the standpoint of a concerned public (Dewey 1927/1970), instead of giving in to the quest for certainty so characteristic of mathematical and modernist interpretations of proof (and solid expertise) (Toulmin 1992). For a constructivist interpretation of mathematics that recognises the constitutive limits of mathematical truth-claims see the contribution of Van Bendegem, Comijn and Fran?ois. For a further elaboration of the way law deals with the relationship between lay persons and experts, see the contribution of Gutwirth about the jury (also Hildebrandt/Gutwirth/Schreurs 2004).

Preuve and ?preuve

If preuve is translated with proof, ?preuve can be translated with test. This could mean they refer to different phases of a process: ?preuve referring to the testing, investigation or examination of a proposition (for instance: that I am guilty of murdering somebody else), preuve referring to the decision of a judge or jury that this proposition is considered to be correct. Rather than this I would like to look into the historical development of the criminal trial to demonstrate a difference that is more fundamental. The English term to prove derives from the old French term prover, that originates from the latin probare. This meant, amongst others, to test. The related term probus means good. (Merriam-Webster’s online dictionary). So originally it seems that to prove your case meant to test if your case is a good one, not necessarily that your case is established as a good one. Like the term fact the term proof seems to have undergone a profound transformation with the advance of mathematics and the natural sciences, creating the well-known debates on deductive and inductive logic (reasoning from formal axioms or experimental experience). This contribution does not question the importance of mathematical and/or experimental proof, but tries to remind us of the fact that the significance of the truth-claims they entail is not self-evident outside the discipline that produced them (see Van Bendegem/Comijn/Fran?ois on the ecology of proof and argument).

?preuve as contest
In northern Europe, before the emergence of the Frankish monarchies, germanic tribes organised their world without a state or a government. All free men sat down together in what has been called the Thing to decide matters they considered of public interest. Absent a government it was up to them to decide which matters were of public interest and who counted as part of the public. The freedom of these men consisted in the fact that they admitted no authority above them, let alone recognize anyone’s authority to command their actions. The fact that no superior authority was acknowledged did not imply the absence of power relationships. Quite on the contrary: the absence of authority implied that conflicts between or within clans had to be solved by processes of negotiation or mediation, or else revenge would be unavoidable. Revenge or retaliation was a duty rather than a right, and it took quite some convincing to persuade people to refrain from taking violent action. In Frysian times a man that hesitated about revenging the death of his brother, would ask himself: do I want to carry my brother in my purse - when faced with the possibility of accepting compensation. A trial in those times was part of a voluntary jurisdiction, meaning that nobody could be forced to appear; no pre-existing law could be applied; a verdict could not be imposed on parties; the judge was a peer of the parties, his ’authority’ based on personal attributes instead of a hierarchical relationship and, lastly, the judge could not find the law unless parties agreed, so he needed the co-operation of the parties. (Glenn 2004, Immink 1973, Hildebrandt 2002)
With the advance of Frankish monarchy the kings try to establish criminal jurisdiction to get a grip on the lives of their people. Since, however, the relationship between the king and his free men is not entirely hierarchical, the jurisdiction remains largely voluntary, meaning for instance that if parties donot agree to the outcome of the trial, they will enter a judicial duel. The trial is basically a situation that involves risks on both sides, because, as said, there is no institution with a monopoly on violence that can enforce its decision unilaterally. In this setting we find the ?preuve - anachronistically often called a method of proof or evidence - which is basically a contest. This contest is not preparing a verdict, but rather the consequence of a verdict in which the contest is assigned to the parties. Leaving them both to face up to each other as peers, taking a risk with their life, to defend their honor. The base-line of the ?preuve is the ritualised contest, putting both parties at risk and the impossibility to refer to an external authority able to impose its solution on the parties. The ?preuve fits in a society of peers, based on co-ordinate justice.

Enqu?te as investigation: preuve as objective truth
After discussing the medieval ?preuve Foucault moves on to discuss the emergence of a new type of criminal trial, based on an altogether different logic: the continental inquisitorial trial. The heart of this trial is formed by its vertical, hierarchical structure, that deals with two ’parties’: on the one hand the magistrate (who is prosecutor, judge and executioner) and representative of the king and on the other hand the accused, a subject of the king. The crime now is not so much the violation of the honor of a peer, but the offence made to royal authority. As the accused is charged, he is turned into an object of investigation. The enqu?te is part and parcel of a compulsory jurisdiction, jurisdiction shared not by peers but by subjects of the king. The initiative to bring a case to court is now with the government, the accused cannot choose not to appear; a written, substantive law has to be applied; the judge represents the king who can at any time intervene in the proceedings, the relationship between judge and accused are vertical; the judge does not depend on the parties to reach his verdict since he shares the authority of the state and has to apply a pre-existing law; lastly this verdict can be imposed upon the accused. Within this type of trial the risk is entirely with the accused, not with the magistrate (or with the king). The closure is not determined in a contest between peers, but after careful and thorough investigation: inquirere (investigation) is the heart of the inquisitorial trial. The subject of the king is the object of investigation. As several scholars of medieval history have emphasised there is no continuity between the battle of the peers of the early middle ages and the complete subjection of the accused in the compulsory jurisdiction during the royal absolutism of the 16th-18th centuries. The enqu?te calls upon witnesses to testify their relevant experience, where the ?preuve stood for a trial that called upon witnesses to side with a party on the basis of kinship or other loyalties. The closure that follows the enqu?te is based on the idea that truth has been established: preuve or proof indicates objective truth. Foucault even seems to suggest that the objectification of the accused that is produced by the enqu?te, and the search for truth instead of consensus are the germ for the scientific revolution of the 16th/17th century: investigating nature by means of experiments, while finding witnesses to testify to the truth of the findings that are claimed (Foucault 1975, p. 587-588).

Exam as disciplinary practice
Foucault continues his investigations with an analysis of the exam. The exam thus forms yet another meaning of the term test: the exam functions - other than the ?preuve - as a calculation of the distance between a given norm and actual behavior, while the norm itself is also produced by the exam. The exam enables categorization and construction of both groups and individuals, paving the way for an instrumentalist welfare society based on a strong state. As we all know Foucault provided us with a masterly analysis of the advance of both the social sciences and disciplinary practices in the 19th century, entirely intertwined now in the contempory sciences for public policy (Jasanoff: science advisors as policy makers 1990). As in Surveiller et punir Foucault opposes the legal theory of Beccaria and Bentham, who plead reforms of criminal and penal practices on the basis of contract-theory (Beccaria) and utilitarianism (Bentham), with the advance of the prison. Classical Enlightenment theory - according to Foucault - seeks a utilitarian justification for punishment (protection of society against criminal actions) and celebrates the principle of legality, emphasizing the protection against arbitrary punishment by means of positive law. In practice, however, the instrumentality of the criminal law is subverted by what Foucault terms disciplinary practices that are not so much focused on protecting society against crime while protecting citizens against arbitrary punishment, but focused on establishing control and surveillance over the lives of individuals. The instrumentality of the criminal law is thus rendered impotent by the instrumentalist intertwining of criminology and criminal policy; the expert thus rises to high status because he can supply the means to identify dangerous individuals while providing the social technologies to normalize individuals even before they become dangerous.

Fair trial
I do not wish to contend Foucault’s analysis of ?preuve, ?nquete and exam. I think it a brilliant historical exposition of the trial as a contest and as an investigation and a demystifying demonstration of the impotence of classical legalism against the advance of the social sciences. However I would like to move on beyond ?preuve and enqu?te to the trial that emerged after the French Revolution, side by side with the disciplinary practices Foucault discusses.

The trial that developed in the course of the 19th and 20th century, now generally referred to as the ’fair trial’ - in European context the fair trial of art. 6 of the ECHR - is an interesting phenomenon. In other work I have elaborated on the fragile historicity of this fair trial and on the way the constraints of our democratic constitutional state constitute its core. In that framework I have described the opposition between the formal and the substantive conceptions of the ’Rechtsstaat’, opposing a legalistic conception of law to the relational conception of law that goes back to Montesquieu and Beccaria (Hildebrandt 2002:417-499). This fair trial is interesting because of its contraints: it expressly organises an equality of arms between a defendant and the prosecution and it expressly does so by means of a public hearing in which the defendant can contest the accusations, based on the investigations of the prosecution. In a sense the fair trial integrates ?preuve and enqu?te, contest and investigation, while preparing closure on the matter at hand (Hildebrandt 2002:319-417). This is where the trial of the expert comes in.

Deferring to or trying science in court?
As Foucault and others have highlighted, enqu?te or investigation presumes hierarchy and subordination. Historically it can only emerge after the popes and later the kings have seized power and built an effective and efficient bureacracy that can turn the peers of feudal medieval society into subjects (Immink 1973, Berman 1985). Investigation presumes dominance, inquiry implies that the object of inquiry can be controlled in some way, put to the test, examined - subjected into the status of being an object. If we follow Francis Bacon this domination is precisely the core-business of science: a succesfull experiment demonstrates subjection of the non-humans (nature) by the humans (Gutwirth 1993:93). You could say that after humans were subjected by the king, things were subjected by humans (one wonders if it could be a co-incidence that Bacon - who is a lawyer - sides with the royal absolutism of Elisabeth I and James I).

It is interesting also to note that one of the prime characteristics of the enqu?te, the involvement of lay-witnesses that testify about what they have seen, was part and parcel of the experiments of the 17th and 18th centuries. These were performed in the public sphere of salons and academies, calling on lay persons to witness and testify. It is only in the 19th century that science began to retreat into the secluded space of the laboratories, creating - in the course of the 19th and 20th century - enormously complex imbroglios of high-tech instruments and human theory, producing not only whole sets of new facts but also a type of expertise that seems inaccesible for the lay person.

Somewhere in this retreat of science from to salons to the academies (where no women were allowed) and from the academies to the laboratories (where lay persons had no business), a deal seems to have been struck: scientists will take it upon them to represent the world of non-humans (brought under supervision in the laboratory), while politicians will represent humans (brought under supervision as sovereignty was invented). [Latour’s double mirror - self-portrait - of modernity]

Experimentation - in this line of reasoning - developed from investigation (a line of thought I find with Foucault also). Not from the exam, that measures deviance from a norm (while at the same time establishing the norm). Also not from the ?preuve, that implies that both parties take a risk. And also - though I did not elaborate on this, because of lack of time - the ?preuve functions in a world that is sacred, where things cannot be subjected into objects (Glenn 2004, chapter 3). This (the link between investigation and experimentation) to me is the difference between experience and the scientific experiment. In case of an experiment the situation to be tested is brought under control, while experience covers a much broader field not necessarily constrained by artificial means with the purpose of testing a hypothesis. The fact that an experiment is a confrontation with the resistance of non-humans, does not yet make it a contest, I think. The king that investigates his subjects can expect to be resisted, but - like the scientist - he took the initiative, constructed the artificial constraints of the inquisitorial trial (of the laboratary procedure), and - again like the scientist - the king does not put very much at risk. Just think of animal testing, however responsible and careful this may be done, it should be obvious who is in charge here.

Now, back to the fact that the retreat of science into the laboratory established or reinforced a deal between science and politics. In court this line of reasoning leads to a particular idea about the relationship between the judge and the expert: the judge decides the legal matter (after all she is a legal expert) after the expert has advised her on the facts. The judge defers to the expert (Jasanoff 1995: 61-63). This seems a rather convenient position for both the lawyers and the scientists. The lawyers can immerse themselves in their legal expertise, trusting the scientists that can claim their truth of the matter. However, as we indicated in the beginning of the paper, expertise has become an essentially contested concept (Gallie 1956): multiplicities of incompatible expert opinion together with questions of relevance that cannot be answered by the consulted expert alone, have created a situation in which expertise is put to the test. In court. The decision on its validity and relevance ultimately resting with judge or jury - both lay persons as far as most of the scientific evidence is concerned. This implies that the results of the investigation of the expert - and even the investigation itself - can and often will be contested in court. Not only on grounds internal to the discipline that produced the expertise, but also on grounds related to the matter of concern that has to be decided in court. The beauty of a court decision is that it performs closure on a matter of concern by establishing the facts in direct relation to the relevant legal norm (factum est). While in a scientific context closure on facts will be postponed untill serious consensus is reached, in a legal context closure will have to take place because of the need for legal certainty. At the same time, we can say that while in a scientific context the black box of the constructed fact can always be opened, in the legal context the closure will at a certain point be definite. For that is what law does for us: produce a stable point of reference from where we can get on with our lives. That is why - in the end - a court decision is about a matter of concern, integrating the establishment of fact with the relevant legal norm, thereby re-establishing the normative force of the legal norm.

The setting of the fair trial, as indicated earlier, organises a space in which contest and investigation are integrated. On the one hand the criminal trial confirms the asymmetrical relationship between the state (prosecution) and the citizen (defendant). This provides the possiblity of closure. On the other hand it provides the legal instruments to contest the accusation of the state. This double instrumentality is the most salient constitutive feature of the substantial conception of the ’Rechtsstaat’: law as an instrument of protection against crime (based on the monopoly on violence), while the same law also protects against use and abuse of this monopoly. The law is thus instrumental for those that oppose it. We call this the paradox of the ’Rechtsstaat’.

This paradox also works for the testing of expertise. Instead of deferring to the expert, the court will consult the expert and - if good reason arises - arrange for it to be contested. Or, in the perspective of the adversarial trial, parties will bring in their experts and see to it that the expertise of the other party is put on trial. After that it is not the expert that decides the matter of fact but the judge or the jury, it is also not the case that we have to wait for the experts to be sure of their case: the presumption of innocence implies that in case of doubt the defendant will have to be acquited.

Conclusion: the trial of the expert, ?preuve and preuve

Traditionally (in our Western, rational tradition) preuve is understood as proof (closure) of a matter of fact, while ?preuve indicates the risky contest over a matter of concern, often judged to be highly irrational.
In this paper I have taken another perspective, by following the complexities of co-ordinate and sub-ordinate justice and their specific integration in the fair trial of the democratic constitutional state (Rechtsstaat). The ?preuve - a contest between free men - part of a voluntary jurisdiction, implies risk on both sides; preuve is the outcome of an enqu?te or investigation by a sovereign on one of his subjects, part of a compulsory jurisdiction, implying the risk lies with the object of investigation. The integration of sub- and co-ordinate justice opens new perspectives on the testing of scientific expertise: after the black box is constructed in the laboratory, as a result of succesfull experimental investigation, it can be contested by those who will suffer or enjoy the consequences in a particular case. This may lead to the production of new scientific knowledge - spurred by the questions of lay persons whose rights and freedoms are at stake. Jasanoff - in Science at the Bar - gives a pertinent example of what she describes as the fact that law and science ’are in fact mutually constitutive in ways that have previously escaped systematic analysis’ claiming that both institutions ’jointly produce our social and scientific knowledge, and our relationships with technological objects’. The example she gives concerns genetic profiling. Originally the expertise in this field was generated by molecular genetists and molecular biologists. After court cases in which identification on the basis of DNA-fingerprinting was challenged, a debate followed in Nature about the methodologies used. Population genetics entered the field, producing new insights into the construction of the gene. In fact the debate on DNA-fingerprinting has led to an awareness of the fallability of ordinary fingerprints, a forensic expertise that - unjustly - gave categorical opinions as regards identification. In England and in the Netherlands recent cases demonstrated how both the validity and the relevance of statistical expertise can be challenged, evoking highly technical discussions on baysian and other types of statistics - in court, in dayly newspapers and in professional literature. The pungency of the questions that lay persons (judges, council for the defense, prosecutors and juries) articulate to test the expertise put before them clearly demonstrates both the capacity of publics to penetrate the validity and relevance of expertise and the fact that law and science are on many occasions mutually constitutive.

Literature:

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

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

W.B. Gallie, Essentially Contested Concepts, Proc. Aristotelian Society 1955-6 (51)

H. Patrick Glenn, Legal Traditions of the World (second edition), Oxford: Oxford University Press 2004

M. Hildebrandt, Straf(begrip) en procesbeginsel, Kluwer Juridische Uitgevers: Deventer 2004

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

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

P.W.A. Immink, La libert? et la peine. ?tude sur la transformation de la libert? et sur le d?veloppement du droit p?nal public en Occident avant le XIIe si?cle, Assen: Van Gorcum 1973

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

Barbara J. Shapiro, A Culture of Fact. England, 1550-1720, Ithaca and London: Cornell University Press 2000

Stephen Toulmin, Cosmopolis: The Hidden Agenda of Modernity, Chicago: Chicago University Press 1992