Rethinking the Standards of Proof

Author(s): Michele Taruffo

Source: The American Journal of Comparative Law, Vol. 51, No. 3 (Summer, 2003), pp. 659-677
Published by: American Society of Comparative Law
Stable URL: http://www.jstor.org/stable/3649122
Accessed: 08/03/2010 15:58

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=ascl.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact [email protected].

American Society of Comparative Law is collaborating with JSTOR to digitize, preserve and extend access to
The American Journal of Comparative Law.

http://www.jstor.org
 MICHELETARUFFO

Rethinking the Standards of Proof

A recent article entitled A Comparative View of Standards of
Proof by Kevin Clermont and Emily Sherwin1 has the merit of dis-
cussing an issue that is not frequently analyzed and of proposing
some interesting remarks about several complexand puzzling topics.
Since the authors' intention seems to be also that of starting a de-
bate,2 this is a welcome opportunity,for a civilian lawyer and a com-
parativist to take part.
The Clermontand Sherwin essay is stimulating and provocative.
Their main claim, that in civil litigation courts should apply a stan-
dard of proofbased on the "preponderanceof evidence"or on the "pre-
vailing probability,"is sound. However, I believe--in contrast with
their approach-that the Bayesian interpretationof judicial decision-
making in terms of quantitative probabilityis largely unsatisfactory,
and that models based on logical probabilityand fuzzy logic may pro-
vide better rationalizations of the evaluation of proofs.3
On the other hand, my main impression is that Clermont and
Sherwin's discourse suffers from some substantial flaws. Since I be-
lieve that eliciting gaps and misunderstandingsis an essential part
of the "trialand error"methodthat characterizesthe advancementof
sciences, includingthe legal ones, I shall try to emphasize here what,
to my view, are the main defects of Clermontand Sherwin'sanalysis.
For sake of clarity, and leaving aside a number of other aspects
that could also be disputed and contested, I shall focus my remarks
upon three main points: 1) a reductivist fallacy 2) a lack of adequate
specific information3) a group of theoretical inadequacies.

1. REDUCTIVIST FALLACY

In their comparative analysis Clermont and Sherwin borrow

from the Germanliterature some isolated statements concerningthe
MICHELE TARUFFO is Professor of Law, University of Pavia (Italy).
1. Clermont & Sherwin, "A Comparative View of Standards of Proof," 50 Am. J.
Comp. L. 243 (2002).
2. Id. at 244.
3. I developed the arguments supporting this approach in Michele Taruffo, La
prova dei fatti giuridici (1992), now translated into Spanish with the title La prueba
de los hechos (2002). See the Spanish edition (to which further references will be
made), at 167.

659
660 THE AMERICAN JOURNAL OF COMPARATIVE LAW [Vol. 51

standard of proof that seems to be applied by German courts. They

then however, in fact make France their principal case study, and
proceed to generalize that study as reflective of all civil law systems.4
This is an unwarranted inference that leads to a reductivist fal-
lacy. There is no doubt that France is a very important procedural
system in Continental Europe, but from a comparative point of view
it is incorrect to take the French system as fairly representative of
civil law systems in general, both at the level of the procedural regu-
lation globally considered and at the more specific level concerning
evidence and proof. Clermont and Sherwin seem to share a misunder-
standing that is particularly widespread in the American legal cul-
ture; namely, that there is just one civil law model that could be
properly described just by taking at random one country (in the pre-
sent case, France; though most frequently American scholars prefer
Germany) and extending what is found in that single country to all
the other civil law countries.
Unfortunately for those who adopt such a reductivist method of
comparison, the reality does not support this kind of inference, in
particular when the French system is taken as the starting point.
Once upon a time, it was otherwise: the French code de procedure
civile enacted in 1806 indeed was taken as a model by many procedu-
ral codes of the 19th century, including the German Zivilprozes-
sordnung enacted in 1877 (but not including the Austrian
Zivilprozessordnungen of 1815 and 1895-98, nor the Spanish Ley de
Enjuiciamiento Civil of 1881, nor the Japanese procedural code of
1890). However, the influence of the French model did not last for-
ever; the situation changed substantially in the course of the 20th
century in the train of many general political and legal developments.
Moreover, in the specific area of civil procedure the innovative exam-
ple provided by the Austrian Zivilprozessordnung enacted in 1895-
1898 became very influential and dominated the European procedu-
ral culture (with the exception only of France).
Therefore, a century of complex transformations that occurred in
most civil law systems put to an end the monopolistic role of the
French model and led to a much more diversified situation in which
at least three main different models of civil procedure may be identi-
fied, plus some "mixed"systems that do not properly belong to any of
these basic models.
There is the French model, now based upon the 1976 code de pro-
cedure civile that deeply transformed the 1806 code, and is now fol-
lowed in Belgium and to a limited extent in Italy. There is also the
German model, based upon the 1877 Zivilprozessordnung that was
reformed several times, mainly by the 1976 Beschleunigungsnovelle

4. Id. at 247, 255, 259 and passim.

2003] RETHINKING THE STANDARDS OF PROOF 661

and by a recent statute enacted in 2001, that is followed to various

extents in Austria, Scandinavia, Japan and other countries as well.
Moreover, there is also the new Spanish model provided by the Ley de
Enjuiciamiento Civil, enacted in 2000, which includes many interest-
ing solutions that are partially inspired by the German example but
that are also to a large extent original. In such a situation, taking
France as the representative example of civil law procedural systems
is clearly an unwarranted inference that leads to a reductivist fallacy.
Such an inference is specifically unwarranted, moreover, with
reference to two other topics that are relevant here. One topic is the
system of evidence and proof. Actually the French system is almost
unique in many respects among civil law countries. Making short a
story that should be very long, here are just two examples: 1) art. 10
of the French code, that vests the judge with the power to order on
her own motion any legally admissible evidence, has no equivalent in
any other system (except Belgium), since everywhere the active role
of the court is usually defined in much narrower terms-the German
judge, for instance, cannot call witnesses ex officio; the Italian judge
has a general power to order evidence on her own motion only in la-
bor disputes, but in ordinary cases she has much less of such a discre-
tion; the Spanish judge has no power to collect evidence on her own
motion; 2) the rules limiting the use of oral evidence for the proof of
contracts, that are provided by the French code civil (see arts. 1341-
1348) find a partial correspondence in Italy (see arts. 2721-2727 of
the Italian civil code), but do not find any correspondence in other
systems of civil law.
Another character of French uniqueness deals with the impor-
tance of civil procedure in the context of legal culture. Perhaps Cler-
mont and Sherwin are right in saying that civil procedure is a minor
subject matter in France, although French proceduralists may not be
pleased by this judgment. However, their statement that there is a
"civilian inattention to matters of procedure and evidence"5 is cer-
tainly not true when referring to other European countries such as
Germany, Spain and Italy, nor to most of the other civil law countries
around the world. Some people are even critical of the excessive im-
portance of civil procedure and civil procedure scholars in these coun-
tries. At any rate, various legal journals (some of them with long and
respected scientific traditions) deal exclusively with procedural mat-
ters,6 and dozens of articles on procedural and evidentiary topics are
published every year by other law reviews. In each country there are
dozens of textbooks on civil procedure, plus a number of commenta-

5. Id. at 259.
6. See for instance the Zeitschrift fiir Zivilprozess in Germany, the Rivista di
diritto processuale in Italy, the Revista de derecho procesal in Spain and the Revista
do processo in Brazil, just to quote some among the better known reviews specifically
concerning procedural law.
662 THE AMERICAN JOURNAL OF COMPARATIVELAW [Vol. 51

ries and treatises including the law of evidence, and hundreds of

monographs concerning a wide array of procedural topics and eviden-
tiary matters.7 In such a situation, the inference drawn by Clermont
and Sherwin from the French situation has no support in the real
condition of the procedural literature existing in these other civil law
countries.

2. LACK OF ADEQUATE SPECIFIC INFORMATION

In the economy of Clermont and Sherwin's discourse, an impor-

tant point is that in civil law countries there would be no substantial
research about the standards of proof and that, in particular, there
would be no knowledge and application of the concept of probability
as a key for a rationalization of factual decision-making.8 Once again,
however, Clermont and Sherwin are betrayed by their "France-is-all"
or "France-or-nothing" reductive bias.
It is true that the French literature is not rich in research con-
cerning the standards of proof, although the French landscape may
appear less sparse if one goes back, for instance, to classical little
book written in the Sixties by Levy-Bruhl9 or, considering that a rele-
vant group of "substantive" provisions concerning evidence and proof
are located in the civil code, one were to look for information on this
topic where it is more prevalent; i.e., in the treatises of private law
and specifically in the parts concerning the proof of contracts.10
However, even in this respect the French situation is very pecu-
liar. Actually the topics of the standards of proof and of the applica-
tion of probabilistic models are discussed to a significant extent in
other civil law legal cultures. In Sweden, for instance, the analysis of
probative values according to probabilistic models has been devel-
oped by Per Olof Ekel6f in some well known essays, and by other
scholars as well.1" In Germany, an important book about the free
evaluation of proofs was published in 1979 by a prominent scholar in
the area of civil procedure.12 Moreover, in the German legal litera-
ture there is a broad and interesting discussion developed by several
books, articles and commentaries dealing specifically with the appli-

7. An even summary bibliography would require several pages and any choice
would be arbitrary and reductive. Even a short glance at the reviews quoted supra n.
6 would provide the reader with sufficient information.
8. Id. at 259.
9. See Levy-Bruhl, La preuve judiciaire. Etude de sociologie juridique (1964).
10. See for instance Ghestin & Goubeaux, "Introduction generale," in J. Ghestin
(ed.), Traite de droit civil 447 (1977) .
11. See e.g., Ekelof, "Beweiswurdigung, Beweislast und Beweis des sersten An-
scheins," 75 ZZP 289 (1972); id., "Free Evaluation of Evidence," 8 Scand. St. in Law 5
(1964). See also the essays collected in Evidentiary Value: Philosophical, Judicial and
Psychological Aspects of a Theory, ed. by P. Gardenfors, B. Hanson and N.-E. Sahlin
(1983).
12. See Walter, Freie Beweiswiirdigung (1979).
2003] RETHINKING THE STANDARDS OF PROOF 663

cability of probabilistic frameworks in the rationalization of eviden-

tiary inferences.13 In the Spanish literature the topic of evidence and
proof has long been a traditional subject matter, and even in recent
years some important monographs have been published.14 In Italy
the situation is similar, beginning in 1914 with the classical La prova
civile, written by one of the greatest Italian legal scholars, Francesco
Carnelutti, and including several books and dozens of articles written
by civil and criminal proceduralists.15 If a personal note is allowed,
and although quoting oneself may sound inelegant, I may add that I
gave a contribution to the research in this domain with a book pub-
lished in 1991, the central core of which is an extended discussion
concerning the standards of proof and the use of probabilistic models
in the decision-making about the facts in issue.16
One may well say, but only after due consideration of the ex-
isting large body of literature, that the outcomes of such an extended
discussion are unsatisfactory, disputable or even mistaken, but it
cannot be said that no research has been conducted in civil law coun-
tries about the subject matter in which Clermont and Sherwin are
interested.
Glancing through the abundant civil law literature on this mat-
ter one may find out that the use of probabilistic calculi is broadly
discussed but is considered with some skepticism, although there are
also recent opinions supporting the application of statistical or quan-
titative probability to the problems of proof.17 At any rate, in Europe
there is no "Bayesian orthodoxy" comparable to that that is optimisti-
cally said to exist in some areas of American legal theory.18 The rea-
sons for the prevailing skeptical orientation that is widely shared in
civil law countries are not the result of ignorance or difficulty to un-
derstand the problem (as Clermont and Sherwin seem to believe),19

13. Just to quote some of the monographs dealing with the topic: Greger, Beweis
und Wahrscheinlichkeit (1978); Leipold, Beweismass und Beweislast im Zivilprozess
(1985); Maassen, Beweismassprobleme im Schadenersatzprozess (1976); Motsch, Vom
rechtsgenigenden Beweis (1983); Nell, Wahrscheinlichkeitsurteile imjuristischen Ent-
scheidungen (1983); Nierhaus, Beweismass und Beweislast (1989); Schneider, Beweis
und Beweiswurdigung (4. Aufl., 1987); Schreiber, Theorie des Beweiswertes fur Bewe-
ismittel im Zivilprozess (1968); Musielak-Stadler, Grundfragen des Beweisrechts 63
(1984).
14. See e.g. Gascon Abellan, Los hechos en el derecho. Bases argumentales de la
prueba (1999); Garcimartin Montero, El Objecto de la Prueba en el Proceso civil
(1997); Cabanas, La valoracion de las pruebas y su control en el proceso civil (1992);
Igartua, Valoracion de la prueba, motivacion y control en el proceso penal (1995).
15. See Carnelutti, La prova civile (1914), and in the recent literature Lombardo,
La prova giudiziale. Contributo alla teoria del giudizio di fatto nel processo (1999).
16. See Taruffo, supra n. 3.
17. See e.g. Garbolino, "Explaining Relevance," 22 Cardozo L.Rev 1503 (2001).
18. See Edwards, "Summing Up: The Society of Bayesian Trial Lawyers," in
Probability and Inference in the Law of Evidence. The Uses and Limits of Bayesian-
ism, ed. by P. Tillers and E.D. Green (1988) at 337.
19. See id., at 260.
664 THE AMERICAN JOURNAL OF COMPARATIVELAW [Vol. 51

but by a widespread and possibly well-founded theoretical skepticism

about the all-encompassing power of statistical probability. Such a
skepticism is supported by several arguments that cannot be dis-
cussed here; they include -inter alia- the availability of different and
more adequate approaches based upon logical probability20 and the
awareness of the extreme complexity of the task of fashioning satis-
factory logical models for factual decision-making. These reasons may
or may not be accepted, but the range of the possible approaches to
the problem of proof is much wider than a simple yes-no answer to
the question whether the application of the Bayes' theorem offers re-
liable solutions for the rationalization of judicial decision-making.21
There is no doubt that various probability theories may offer in-
teresting insights. However, probability is neither a simple concept
nor a ready-made and catch-all device applicable to any kind of issue.
On the other hand, even in the Anglo-American literature the Baye-
sian approach is not generally shared22, and other interesting models
are proposed for the logical analysis of probative inferences23. There-
fore, on both sides of the "America vs. rest-of-the-world" divide things
are very complex and diversified.

3. SOME THEORETICALINADEQUACIES

Clermont and Sherwin's main thesis is that civil law countries

apply the same standards of proof in criminal and in civil cases, and
that therefore they apply the particularly high standard of "proofbe-
yond reasonable doubt" also in deciding civil cases. However, this pic-
ture of the practice of civil law courts in making decisions about
proofs seems to be largely unwarranted from several points of view.

a) First of all, it has to be stressed that Clermont and Sherwin

do not offer enough preponderant evidence (let alone a proof beyond
reasonable doubt) supporting their claim. They adduce some scat-
tered statements made by some European scholars (mainly French),
taking them out of their theoretical context and without checking to
which extent they correctly represent the practice of courts in civil

20. See mainly Cohen, L.J., The Probable and the Provable (1977).
21. For a presentation of various reasons justifying the rejection of the Bayesian
theory of proof see e.g., Taruffo, supra n. 3, at 193. See also id. at 223, 241, for the
description of alternative models of the evaluation of proofs in terms of logical
probability.
22. See e.g., Brilmayer, "Second-order Evidence and Bayesian Logic,", in
Probability and Inference, supra n. 16, at 164: Callen, "Notes on A Grand Illusion:
Some Limits on the Use of Bayesian Theory in Evidence Law," 57 Ind.L.J. 3 (1982).
23. See e.g., Anderson and Twining, Analysis of Evidence: How to Do Things with
Facts Based on Wigmore'sScience of Judicial Proof 329 (1991), proposing an updated
version of Wigmore's chart, and Schum, Evidence and Inference for the Intelligence
Analyst (1987).
2003] RETHINKING THE STANDARDS OF PROOF 665

law countries.24 Clermont and Sherwin have no empirical evidence to

present about the real application that civil law courts make of any
standards of proof; it seems evident that such evidence would be
needed in order to describe and criticize the practice of those courts.
As to the normative sources, it has to be emphasized that no rule,
in any civil law system, requires the courts to apply in civil cases the
same standard of proof that is applied in criminal cases. Indeed, the
standard of proof beyond reasonable doubt is not explicitly stated
even in criminal cases: no wonder, then, if it is neither stated nor
generally applied in civil cases.
With reference to the French legal literature, Clermont and
Sherwin are right in stressing that in France the handbooks of civil
procedure do not even speak of standards of proof. This lack may be a
matter for criticism, but it is hardly proof that French courts apply
the standard of the "proofbeyond reasonable doubt". It is worth men-
tioning that such a standard is not provided and is not applied even
in criminal procedure.25
So far as to the Italian system is concerned, nowhere in the legal
literature it is said that civil courts apply such a standard, for the
good reason that it is not true.26 As to criminal courts, in discussing
the evaluation of proofs a well-known expert on the topics of criminal
evidence and proof makes no reference to any standard of proof be-
yond reasonable doubt.27 Moreover, a leading authority in criminal
law-and a prominent practitioner-strongly criticizes the Italian
criminal courts for not applying the standard of proof beyond reason-
able doubt, and for relying upon the much lower standard, that is
typical of civil cases, of the prevailing probability.28 Just recently the
Italian Supreme Court seems to have started a revirement in the

24. See id. at 247, 254.

25. See e.g., a well known handbook of criminal procedure (Stefani-Levasseur-
Bouloc, Procedure Penale, XIII ed. 45 (1987), where it is emphasized that the judge is
completely free to evaluate the probative force of proofs, without giving any reason of
her judgment. This means that actually the judge is not required to apply any stan-
dard of proof, since she decides only according with her personal conscience. This
seems just the contrary of the application of a high standard of proof such as that of
the proof beyond reasonable doubt.
26. A recent overview of the Italian case law concerning this topic shows that
Italian courts in deciding civil cases adopt very flexible standards of proof, based es-
sentially upon the discretion of the judge, without any reference to the standard of
proof beyond reasonable doubt. See Commentario breve al codice di procedura civile.
Complemento giurisprudenziale. Ed. by F.Carpi and M.Taruffo 477 (3rd ed. 2002).
27. See Ubertis, Argomenti di procedura penale 136, 139 (2002), where the focus
is upon the persuasive effect of proofs and the discretion of the court, but no specific
standard is determined.
28. See Stella, Giustizia e modernitd 147, 328 (2nd ed. 2002).
666 THE AMERICAN JOURNAL OF COMPARATIVELAW [Vol. 51

sense of adopting in some criminal cases a standard that is substan-

tially equivalent to that of the proof beyond reasonable doubt.29
In fact, even assuming -notwithstanding the lack of empirical
evidence- that there is no clear distinction between civil and criminal
standards of proof, this does not imply that the civil standard is
equivalent to that of the "proofbeyond reasonable doubt". Rather-as
Clermont and Sherwin seem to understand30-where there is such
an equivalence it means that criminal courts apply the lower stan-
dard of the preponderance of evidence that is typical of civil courts,
not that civil courts apply the higher standard of proof beyond rea-
sonable doubt.

b) Much of Clermont and Sherwin's misunderstanding derives

also from some disputable theoretical assumptions that deserve to be
mentioned.
One of these inadequacies -perhaps the most important- con-
cerns just the core concept of the standard of proof that is applied in
civil law systems. Clermont and Sherwin assume that the French
standard of the intime conviction is equivalent to the standard of
"proof beyond reasonable doubt", and that it is used as such in all
civil law systems to decide civil cases. Both these claims are
unfounded.
The principle of the intime conviction and all the similar (but not
identical) principles concerning the free evaluation of proofs do not by
themselves entail the adoption of any specific standard of proof, let
alone the standard of proof beyond reasonable doubt. The history of
these principles, as well as their systematic role in modern systems,
show that they have a negative rather than a positive meaning. Their
negative meaning is that to the extent they are applied (which is in
many cases a matter of degree) they exclude the application of rules
of legal proof (i.e., rules determining in general and binding terms the
probative force of specific items of evidence), vesting the court with
the power to determine the weight of proofs on the basis of a discre-
tionary evaluation. In a word: these principles exclude the applica-
tion of legal standards of proof but do not prescribe by themselves
any positive standard of proof.31
The real problem, therefore, is to determine how the "empty
space" created by the principles of free evaluation is "filled up" by the
practice of courts and by criteria suggested by legal theorists. We do
not know much about the actual practice of courts because of the lack
of empirical research in this domain, and thus we have to rely upon

29. See Stella, Etica e razionalitd del processo penale nella recente sentenza sulla
causalita delle Sezioni Unite della Suprema Corte di Cassazione, in XLV Rivista ital-
iana di diritto e procedura penale 767 (2002).
30. Id at 262.
31. About these topics see, for a broader analysis, Taruffo, supra n. 3, at 394.
 2003] RETHINKING THE STANDARDS OF PROOF 667
what legal theoristssay, althoughsometimesit is not clearwhether
theirstatementsareprescriptive(givingsuggestionsto the courts)or
descriptive(picturingwhat courtsreallydo),or both.
In this perspectiveit is worthemphasizingthat the principleof
freeevaluationis statedin differenttermsin the varioussystems.In
Francethe traditionalformulaof the intimeconviction(that dates
back to the FrenchRevolutionbut is not stated anywherein civil
codes)stresses the value of the subjective"intimate"persuasionof
the singlejudge,relyingmainlyuponher individualand even emo-
tionalbeliefs.It does not mean by itself, however,that such beliefs
must be based upon a particularlyhigh standardof proof:what is
requiredis that thejudgebe "intimately" andindividuallypersuaded
of something32Roughlyspeaking,it does not matter whetherthe
judgehas speciallystrongrationalandevidentiarybases(in termsof
probability)to believewhat she believes.Herdeepindividualconvic-
tionhas to be well-rootedin her feelings,in orderto producea "moral
certainty"or a preuve morale aboutthe facts of the case. This does
not meanrelianceuponany clearstandardof proof:it meansto rely
just uponthe personalconscienceof thejudge.Thisconceptionof the
evaluationof proofsmay well be stated in terms of"truth","moral
certainty"or alike,but these areidealizedandwishful-thinking ways
to definethe final outcomeof the decisionon the facts, ratherthan
clear definitionsof the standardsthat are or shouldbe appliedby
courts.In a word:the intimateconvictionof the judge is a sort of
blackboxthat is left to the personalconscienceofthejudge,whilethe
judgmentis expressedusing the rhetoricof"certainty"and "truth"
notwithstanding the factthat thejudgeis not requiredto explainthe
reasonssupportingher conclusions.
Othersystems do not followthis subjectivisticperspective.For
instance,the freie Beweiswurdigungprinciplestatedby § 286 of the
GermanZivilprozessordnungstressesthe freedomof the courtin the
evaluationof proofswithoutany subjectiveovertone.33 Thesamerule
says also that the courtshoulduse its discretionin orderto deter-
minewhetherthe facts of the case are true or false. This statement
has raised, in Germanlegal theory, the problemof the so-called
Beweismasse,i.e., of the degreeof proofthat is consideredas neces-
saryin orderto say that a factis "true".Yearsagosomeauthorspro-
posedthe standardof the uberwiegendeWahrscheinlichkeit,34 that is
of the preponderantprobability,but the prevailingopinionseems
nowto be that sucha standardis toolowto fit with the ideaof estab-

32. See Stefani-Levasseur-Bouloc, supra n. 20.

33. On the contrary,the applicationof objectivestandardsis clearlystressed. See
Prutting,in MunchenerKommentarzur Zivilprozessordung(1992), vol 1., at 1704,
1708.
34. See the literature quotedsupra n. 13.
668 THE AMERICAN JOURNAL OF COMPARATIVELAW [Vol. 51

lishing the truth of the facts that is at the basis of § 286.35The Bewe-
ismasse problemraised a discussion that cannot be summarizedhere:
suff1ceit to say that the probabilistictheory of the standards of proof,
albeit proposed by several influential scholars, has been rejected
mainly because it was based upon conceptionsof quantitative or sta-
tistical probabilitythat were consideredas not applicableto judicial
reasoning about facts. Even when the vocabulary of probability is
still used, it is only in orderto exclude any referenceto ideas of"abso-
lute certainty"that would be improperin a judicial context. Corre-
spondingly,the "truth"requiredby § 286 is defined in the general but
vague term of "high probability" (hohe Wahrscheinlichkeit).36
Whether or not this idea of 4'highprobability"may be considered as
equivalent to the standard of proof beyond reasonable doubt is a
problemthat should be discussed on the basis of a thorough analysis
of the German theory, which of course cannot be made here. Some-
times such an equivalence is maintained by Germancourts, but the
views proposedby the German legal theory are much more complex
and sophisticatedand do not take the conceptof proofbeyondreason-
able doubt to be a viable general standard for the evaluation of
proofs.37Whetheror not the prevailing Germantheory is worth shar-
ing in other countries, where different rules concerningthe evalua-
tion of proofs are provided, is a further interesting question that
would deserre a thorough comparative and theoretical discussion
that cannot be made in these notes.
In other civil law systems matters are differentyet. In Spain, for
instance, the traditional reference to the reglas de la sana critzea
(that is repeatedly made by arts. 316 n.2, 348 and 376 of the 2000 Ley
de EnjuiciamientoCivil) is usually interpreted as an exclusion of le-
gal proofs and as a reference to commonsenserules of reason and to
the average experience of the world.38In a similar way, the Italian
general principleof the prudente apprezzamento(stated by art.ll6 of
the Italian code of civil procedure)is usually intended as a suggestion
to the courts to be cautious in their discretionary appreciation of
proofs.In general the ground for a rational evaluation of proofis not
defined by reference to formal standards, but by requiring the court
to articulate sound arguments based upon available and reliable
backgroundknowledge (the Ehrfahrungssatzeof the German litera-
ture or the massime d'esperienzaof the Italian legal theory) and to
the standards of inference providedby the commonculture.39Ratio-

35. See Prutting,supra n. 33, at 1705, for a synthetic but thoroughdiscussionof

the problem.
36. See Prutting, supra n. 33, at 1711.
37. See e.g. Musielakand Stadler, supra n. 13, at 69, 74, 77.
38. See generallythe literature quotedsupra n. 14.
39. On this importanttopicsee e.g., Taruffo,supra n. 3, at 208, 219, 334, 509, and
Musielak-Stadler,supra n. 13, at 70.
2003] RETHINKING THE STANDARDS OF PROOF 669

nalizing these references in terms of probability, one should think of

a standard of prevailing probability, certainly not of a standard of
proof beyond reasonable doubt. At any rate, no specific standards of
proof are prescribed to Spanish and Italian courts in civil cases, al-
though they are also supposed to establish the "truth" of the facts in
issue. Correspondingly, it not correct to say that civil law systems
state articulated general standards of proof but do not apply them.40
Actually these systems are not hypocritical: although courts may be
inclined to explain their judgments in idealized or even misleading
terms, the fact is that where a standard is fixed, as in Germany, it is
followed, while in other countries, such as France, Italy and Spain,
there simply are no fixed standards of proof, since the evaluation of
proofs is left to the free discretion of the judge.

c) Another theoretical inadequacy of Clermont and Sherwin's

discourse, which may explain at least in part their difficulty in under-
standing what happens in civil law courts, has to do with the ideas of
probability and judicial truth themselves. They rely upon a rather
naive idea of probability, roughly corresponding to the popular con-
cept of statistical or quantitative probability, and upon a naive and
unqualified idea of truth. Correspondingly, they are in trouble when
they try to define how proofs are or should be evaluated, since simpli-
fied commonsense does not lead very far in this domain, and much
more sophisticated concepts are required in order to analyze the eval-
uation of proofs.
It is true that the Bayesian version of the probability calculus is
rather popular in the American theory of evidence, and that there are
even supporters of the use of "naked statistics" as evidence. However,
one should never forget the devastating analysis developed by Nis-
bett and Ross about the mistakes that ordinary people -and judges
are ordinary people from this point of view-make as "quantifiers" of
ordinary experience.41 On the other hand, even within the English-
speaking area other concepts of probability, as for instance the logical
"provability" studied by Jonathan Cohen,42 have been proposed as
means for the rationalization of factual decision-making. Moreover,
other logical tools such as the so-called "fuzzy logic," may be useful in
the attempt to provide courts with rational frameworks for the evalu-
ation of proofs.43 An acritical reference to vague ideas of quantitative
probability, as for instance to the idea that probability values should
be included between 0 and 1 and should be expressed in numbers
such as 0.25, 0.75, and so forth, is generally considered as a very poor

40. See Clermont & Sherwin, at 261.

41. See Nisbett and Ross, Human Inference: Strategies and Shortcomings of So-
cial Judgment (1980).
42. See Cohen, supra n. 20, at 13.
43. Also about this point see Taruffo, supra n. 3, at 228.
 AMERICAN JOURNAL OF COMPARATIVELAW
THE 51
[Vol.
670

analytical tool.It is also misleading,becauseit makespeoplebelieve in

thatthe degreesof probativevalues may be properlyexpressed
simple numbers.44 At anyrate,the questionwhetherprobability,and
kindof probability,maybe used in the evaluationof proofs, is
which are several
stillan open and extremely complicated puzzle. There
good reasonswhythe Bayesianversionof the determination of proba-
versions, may
tivevalues, as well as otherquantitative-probabilistic of the
appear attractivebut are in fact unreliable as rationalizations
reasoningaboutthe facts. Then, if standards as those of the
judicial reasonable
"preponderance of evidence"and of the "proof beyond
are def1ned-as they oftenare-in termsof numericalvalues of
doubt" that the for-
quantitative probabilitywithinthe range from 0 to 1 (so
mercorrespondsat least to 0.51 and the latter to something higher
than0.90),suchdefinitionscannotbe taken as reliable and adequate
rationalizations of the standardsof proof.45
In the domainof evidenceand proofit is frequentto mentionedfind state-
mentsaccordingto which as it is said in the already
§286 of the GermanZivilprozessordnungcourtsshouldestablish
thetruthofthe factsin issue.However,defining"judicial truth"is far
froman easy task. On the one hand,and even if one does not share
the skeptical postmoderntheories claiming that "truth" is non-
of the facts in
sense,46determiningwhat a truthfulreconstruction problem.
issueshouldbe is a puzzlingtheoreticaland philosophical but the
Obviously in judicialcontextsthere are no "absolute" truths,
ideasof "relative", "contextual" and "situated"truths that are sup-
posedto be typicalofjudicialdecisionsare far frombeingintuitivelyuse of any
clear.Onepointthat is clearenough,however,is that the clarify-
unqualifiednotionof truthends up by confusingratherthan whichever
ingthe termsof the problem.Anotherclearpointis that, any refer-
ideaofjudicialtruthis adopted,it doesnot implyby itself
enceto specificstandardsof proof.
A recent book concerningthe philosophicaldimensionof the
problemproposesa definitionaccordingto whichin judicialcontexts It may
onlythe facts that are provedmay be consideredas "true".47 sets aside
soundbanal,but it is not. On the one hand, this definition

44. This is an exampleof the syndromeof "whatcannot be counteddoes not exist"

Mathematics: Precision and Ritual in the
that has been analyzedby Tribe,"Trialby
Legal Process," 84 Harv. L. Rev. 1361 (1971).
standardof "proofbeyondrea-
45. In particular,the quantitativedefinitionof theat
sonabledoubt"is questioned:see Cohen, supra n. 20, 82; id., "TheRole of Eviden-
tial Weightin CriminalProof,"in Probabilityand Inference, supran. 22, at 113:Kaye,
Need a Calculus of Weight to Understand Proof Beyond ReasonableDoubt?,"
"DoWe
id. at 129. Prefciceto a Prag-
46. See e.g., Stephen P. Stich, The Fragmentationof Reason:
matic Theoryof CognitiveEvaluation(1990);Richard Rorty,Consequencesof Pragma-
made by Haack
tism (1982), but see also the sharp rebuttal of these theories
Manifestoof a Passionate Moderate 7 (1998).
47. See FerrerBeltran,Pruebay verdaden el derecho61 (2002).
 2003] RETHINKING THE STANDARDS OF PROOF
671
any improperidea of judicialtruth. In particular,it is a rebuttalof
any conceptionevoking"absolute," "certain"or "indisputable"truths.
Onthe otherhand,it makesclearthat in judicialcontexts"truth"is
nothingbut a functionof the availableevidence.However,this defini-
tion doesnot requireus to connect"truth" with any specificstandard
of proof.Also in the commondiscourse,when it is said that a (civil)
courtis expectedto findout the "truthof the facts",it doesnot mean
that truthdependsuponthe adoptionof any specificandparticularly
high standardof proof.
All this meansthat a judicialtruthmaywell be achievedon the
basis of a preponderantevidence, or of a prevailing logical
probability,concerningthe statementof a fact, at least insofaras
there are no specificprovisionsrequiringa higherstandard(as the
Germantheoristsseem to consider§ 286 of the Zivilprozessordnung
to require).48Actuallythis happensin manycases:by assessingcon-
flictingor contradictory proofs,the courtmakes a balancedevalua-
tion and ends up makinga choicein favor of the relativelymore
reliableversionof the facts.This versionmay be formallypresented
as "thetruth"determinedby the court,but this is a matterof style
and of languageused by judges in writingtheir opinionsand not-
onceagain a matterof applyinghigheror lowerstandardsof proof.
Judgesmaybe pushedto use the languageof truthby the prevailing
ideologyof their role, but the rhetoricused in draftingjudgments
shouldnot be confusedwith the realityofjudicialdecision-making.
Clermontand Sherwin49seem to be amazedby the statement
that in civillaw systems(andspecificallyin France)decisionsshould
notbe "doubtful". However,sucha statementdoesnotmeanthat only
"certain" or "absolute"truthsshouldbe establishedby courtsor that
"truth"impliesa proofbeyondreasonabledoubt.It meansonly that
in writingher opiniona judge will not say: "I am in doubtabout
whetherF is true or false, but neverthelessI decideF".Rather,the
judgewill say "sincethere is strongerevidencein favorof F rather
than againstF, F maybe taken as true, and thereforeI decideF".It
shouldnot soundso strange.Perhapsalso an Americanjudge,if re-
quiredto explaina similardecision,wouldnot say: "I am in doubt
aboutF, but neverthelessI decideF".Thatis to say that in the exam-
ple discussedby Clermontand Sherwinconcerningthe presenceof
conflictingevidenceof the executionof a note,50a civilcourtwill find
in favorof the partywho offereda prevailingevidencein supportof
her positionand will rejectany claimor defensebaseduponinsuffi-
cientevidence.Sucha courtmay be inclinedto explainits judgment
bysayingthat the versionthat was supportedby a preponderant evi-
48. See Prutting, supra n. 33, at 1711.
49. See id. at 272.
50. See id. at 246, 265, 268.
672 THE AMERICAN JOURNAL OF COMPARATIVELAW [Vol. 51

dence was found to be true, but this does not mean that the court did
not make any choice about the relative weight of the conflicting
evidence.
In a word: many civil law systems apply (although with some ex-
ceptions) the principle of the free evaluation of proofs without requir-
ing any reference to especially high standards of proof in civil
litigation. This can be explained or even criticized in several ways,
saying for instance that too much discretion is left to courts, but it
cannot be said that civil law courts always apply the standard of
proof beyond reasonable doubt.

d) In order to support their claim, Clermont and Sherwin use an

argument based upon the rules that in civil law systems govern the
burden of proof.51 However, this argument, too, appears to be sub-
stantially inconclusive. They, again, start from an exceedingly simpli-
fied picture of an extremely complex problem. They are right in
saying that there are general rules concerning the burden of proof
(although, as for instance in Germany, there are no general princi-
ples explicitly stated in the codes, and general standards are induc-
tively derived from a number of specific rules),52 and that detailed
rules are enacted in order to manipulate the allocation of the burden
between the parties, sometimes with the aim of favoring the procedu-
ral position of particular subjects (as for instance that of the plaintiff-
worker in labor disputes). They are also right in saying that the bur-
den of proof is usually regulated by strict statutory provisions, while
American courts apply more flexible and discretionary standards in
shifting the burden between the parties. However, it should be noted
that German and Italian courts are also able to carve out significant
areas of discretion in performing a similar (although not an identical)
operation.53
Nonetheless, all this can hardly be considered as an argument
showing that civil law courts apply a standard of proof beyond rea-
sonable doubt. What Clermont and Sherwin do not consider, in fact,
is the fundamental function that the rules about the burden of proof
actually perform in civil law systems. On the one hand, one should
take into account that in these systems there is no burden of produc-
ing evidence but only a burden of proof or of persuasion. There is a
significant difference between vesting a court with a discretionary
power to allocate the burden of producing evidence, and vesting a
court with the power of shifting from one party to another party the
burden of proof in a strict sense. In civil law countries this burden is

51. Id. at 248, 272.

52. See Pritting, supra n. 33, at 1720, for a synthetic description of the German
situation, and for broad bibliographical information.
53. See e.g., Taruffo, "Onere della prova," in Digesto delle Discipline Privatistiche.
Sezione Civile, vol. XIII 65 (1995); Prutting, supra n. 33, at 1712.
2003] RETHINKING THE STANDARDS OF PROOF 673

usually considered a matter of substantive law, not one matter of pro-

cedural law, with different consequences about the role and the pow-
ers of the court.
On the other hand, although rules concerning the burden of proof
are often stated in positive terms, saying who should prove what
(such as: "plaintiff will prove the facts representing the basis of her
claim"), in fact they are standards for the allocation, between the par-
ties, of the negative effects of the final lack of proof of material
facts.54 In other words, they are aimed at providing the court, which
must always decide the case and may not end up with a non liquet
decision, with criteria that the court is to apply in its final decision-
making, at the point at which it finds out that some material facts
have not been proved. The party that had the burden of proving such
facts, but was not able to prove them, will lose.
This being, in its essential features, the basic mechanism of the
burden of proof in civil law systems, a rather obvious remark is that
it has nothing to do, in itself, with the problem of the standards of
proof. However, a situation that is frequent in the daily practice of
civil law courts (as well as of any court in any system) is that in which
the court has to make a choice among two or more competing versions
of the facts in issue, each version being supported by some evidence
and none being supported by certain proofs. In such a situation the
court cannot but choose the "relatively more probable" version of the
facts, and may reject the claim on the basis of the rules about the
burden of proof only when a material fact has not been proved at all,
or when the evidence about this fact is so weak that it cannot be con-
sidered as "true". In short, rules concerning the allocation of burdens
of proof should not be confused with the standards for determining
the weight of proof that is required in order to consider a fact as
proved.

4. PROCEDURAL SYSTEMS IN SEARCH OF TRUTH?

Clermont and Sherwin's article includes several other points

that would deserve careful consideration, but dealing with them
would take these short notes too far. One more topic, however, is
worth discussing here.
Clermont and Sherwin say that civil law systems are not espe-
cially oriented to the search of truth because they are aimed at the
solution of disputes, while the American procedural system is ori-
ented to the search of truth about the facts in issue.55 Both these
statements may be challenged.

54. See Taruffo, supra n. 3, at 503; id. supra n. 53; Prutting, supra n. 33, at 1721.
55. Id. at 270, 271.
674 THE AMERICAN JOURNAL OF COMPARATIVELAW [Vol. 51

On the one hand, it is reasonable to assume that all procedural

systems are basically aimed at the resolution of civil disputes, and
that this function cannot be considered a peculiar feature of litigation
in civil law systems. It is hard to believe that common law systems
are typically not oriented to the solution of disputes, as Clermont and
Sherwin seem to imply by saying that civil law systems are different
just because they are oriented to solve disputes. This point seems to
need no further demonstration. At any rate, it is worth emphasizing
that there is no logical contradiction between the search for truth and
the resolution of disputes, since disputes may well be solved also on
the basis of truthful decisions.
The only relevant problem is that the goal of dispute resolution
may be pursued in different ways, depending upon the structural and
functional characters of the various procedural systems. There are
systems oriented to solve disputes in ways that do not imply the
search of truth; i.e., systems in which the resolution of disputes -no
matter how they are solved- is the only relevant goal,56 and systems
oriented to solve disputes by means of decisions based upon a truth-
ful reconstruction of the facts in issue.57 It is commonly believed that
civil law systems belong to the second group, while the American sys-
tem belongs to the first one, so that the opposite opinion advanced by
Clermont and Sherwin may appear rather odd and in need of a robust
demonstration that they do not provide.
If they are referring to the fact that in civil law systems there are
alternative methods of dispute resolution (such as arbitration, media-
tion, and so forth), their claim becomes even more amazing. Civil law
systems actually use several ADR devices and encourage the parties
to settle their disputes, but there is no doubt that in these systems (at
least in Continental Europe and in Latin America) the main method
for dispute resolution is still to go to a court and have a case decided
on its merits. On the other hand, it is well known that in the Western
legal world the U.S. has been, at least since the Seventies, the home
of ADR, so that if one takes literally the Clermont and Sherwin's
standpoint one should arrive at the conclusion that the American sys-
tem is relatively the least oriented to the search for truth, rather
than the most oriented in that direction.
However, it is very hard to maintain in general terms that civil
law systems are not oriented to the achievement of possibly truthful
decisions while the American system is so oriented. If this claim is
based, as it usually happens and as it seems to be the case with Cler-
mont and Sherwin,58 on the fact that civil law systems do not have

56. See for instance Mirjan Damaska, The Faces of Justice and State Authority. A
Comparative Approach to the Legal Process 97 (1986).
57. See Damaska, supra n. 56, at 160.
58. See id. at 265.
 2003] RETHINKING THE STANDARDS OF PROOF 675

any American-typediscovery,the argumentdoes not hold. Among

commonlaw systems,actually,onlythe U.S. have an American-type
discovery:in Englandthereis andalwayshas beenonlya rathernar-
row discoveryof documents,59and no other commonlaw system
adoptthe Americandiscoveryprocess.Obviouslyit is not enoughto
say that these commonlaw systemsarenot interestedin establishing
the truth of the facts in issue. On the otherhand,it is not true that
civillaw systemsdonothaveanydiscoverydevice.Thesesystemsare
basedupona generalobligationof the partiesto disclosethe relevant
documents:in case of noncompliance a partymay obtainan orderof
the court(a sortof subpoenaducestecum) addressedto the partyor
non-partyin possessionof the docuInent- directingthis subjectto pre-
sent the documentnot previouslydisclosed.VVhatthese systemsdo
not alloware the "fishingexpeditions" into the papersand recordsof
the otherparty,but it is hardto say that allowingsuchexpeditionsis
a necessaryconditionforthe searchof truth.
In generalterms, moreover,it has to be emphasizedthat the
problemof discoverydoes not stand alone and shouldbe discussed
andunderstoodin the broadercontextof the preparatory phaseofthe
proceedings.VVhen,as happens in civil law systems, a strict fact
pleadingrule is applied,60muchwork has to be done by lawyers
beforefilinga claim,andto decidewhetheror not thereare factssuf-
ficientto supportthe claim.This meansthat when a case is started
there is muchless need forthe partiesto go aroundfishingforfacts
and information.Actuallythe "preparatory phases"in civil law sys-
tems are aimedat definingpreciselythe relevantfacts in issue that
are disputed,and at establishingwhichevidenceis relevantto prove
suchfacts,with an activeroleusuallyplayedby the court.This does
not meanthat suchsystemsare not interestedin the searchof truth,
just the opposite.The fact is that they use differentstrategies,al-
thoughno less sophisticatedones, to achievethis goal. To take the
Germanexample,not only is truth the main goal stated by the al-
ready mentioned§ 286 of the Zivilprozessordnung, the search for
truthis the mainreferencepointof the Germanlegaltheoryconcern-
ing the problemsof proof,and the courtplays a centralrole in the
preparatoryphase of the proceeding,elicitingall the relevantfacts
andcorrelatingthe collectionof evidenceto the effectiveproofof such
facts.6l

59. It is interesting to obserarethat the Civil ProcedureRules 1998, enacted in

1999, do not even use the term "discovery"* and regulate only the disclosureand in-
spectionof documents:see Rule 31.
60. See for instance art.163 of the Italian code of civil procedure,art 339 of the
SpanishLeyde enjuiciamientocivil, and art. 56 of the Frenchcodede procedurecivile.
61. About the German preparatory phase see §§ 275 and 276 of the
Zivilprozessordnung.
676 THE AMERICAN JOURNAL OF COMPARATIVELAW [Vol. 51

In general, moreover, the opinion is commonly shared in civil law

countries is that judicial decision-making should be based upon the
relatively best approximation to the truth of the facts.62 Single proce-
dural systems may be more or less well suited to achieve this goal,
but it cannot be said in general that they do not pursue this goal at
all, or that they are less interested in it than in the American system.
Clermont and Sherwin's claim that American civil procedure is
especially oriented to the search of truth may also sound amazing. I
do not dare to take a strong position on this issue, given the danger of
relying upon incomplete information or misunderstandings, but on
the basis of a good deal of American literature I am inclined not to
share Clermont and Sherwin's statement. Choosing at random
among a wide range of arguments conflicting with this statement one
may say that: a) the adversarial theory of litigation that is deeply
embedded in the American procedural system is praised63 or criti-
cized64 (of course from opposite standpoints) for the very reason that
it does not include the search of truth among the main goals of litiga-
tion;65;b) the American philosophical and legal culture, from Rorty to
feminist jurisprudence,66 is by far the culture that is most directly
engaged in demostrating that truth in general, and judicial truth in
particular, is nonsense; c) a wide set of exclusionary rules such as
those of the American law of evidence is hardly compatible with an
effective search of truth; d) one may wonder whether the American
discovery, with its well known abuses, is really aimed at the search of
truth or if its real purpose is not that of exerting pressures on the
adverse party in order to achieve a forced settlement or of fishing for
evidence that is needed to set up a case, although such evidence will
not be presented at trial; e) in spite of Wigmore's optimistic defini-
tion, one may wonder whether the American cross-examination is re-
ally "the greatest legal engine ever invented for the discovery of
truth,"67 or whether it is a "mental duel" in which the lawyer's main
purpose is to "destroy the witness," mainly when she is telling the

62. See generally Taruffo, supra n. 3, at 56.

63. See e.g., Saltzburg, "Lawyers, Clients, and the Adversary System," 37 Mercer
L. Rev. 647, 650 (1986).
64. See e.g., Frankel, "The Search for Truth: An Umpireal View," 123 U. Pa L.
Rev. 1041 (1975); Marvin E. Frankel, Partisan Justice 73 (1980).
65. That the American adversarial system is truth-averse is very clear on the ba-
sis of Damaska's analysis. See supra n. 56, at 123, where he says that in such a sys-
tem the truth "tends to appear elusive and ambiguous". See also infra n. 69.
66. See e.g., Richard Rorty, Truth and Progress. Philosophical Papers. Vol. 3, 1
(1998); Williams, "Feminist Legal Epistemology," 7 Berkl. Women'sL.J. 63 (1992). See
also supra n. 46.
67. See John H. Wigmore, A Treatise on the Anglo-American System of Evidence
in Trials at the Common Law, 3rd ed., vol.V, 29 (1940).
2003] RETHINKING THE STANDARDS OF PROOF 677

truth, as is sometimes suggested to skillful cross-examiners;68 f) inde-

pendently of the political foundation of the jury-trial system, one may
wonder whether a jury verdict -mysterious and inscrutable as it is-
may be considered as an efficient method for the achievement of
truthful decisions (after all, nobody knows how and why a jury actu-
ally reaches a verdict, and therefore nobody knows whether or not the
"truth" has been discovered by jurors). Correspondingly, it is hard to
say whether or not a criminal jury actually applies the standard of
proof beyond reasonable doubt, and whether or not a civil jury actu-
ally applies the standard of the preponderance of evidence.
Each of these issues, and many others, should be thoroughly dis-
cussed and carefully verified before claiming that the American sys-
tem is especially well-suited for the discovery of truth. At any rate,
one should take into consideration the impressive picture that a lead-
ing authority in the American legal culture has recently provided of
the American perception of the problem of judicial truth: "There is no
such thing as objective truth. There is white truth and black truth
and Asian truth and Hispanic truth. There is male truth and female
truth. Probably there is young truth and old truth, and rich truth and
poor truth, too."69 No less significant is Mirian Damaska's sharp
judgment, according to which "the Anglo-American method of collect-
ing and presenting evidence not only deviates from ordinary decision-
making but. .it also strikes discordant notes with arrangements rec-
ommended by a model of inquiry aimed at obtaining only accurate,
trustworthy knowledge."70

68. See e.g., J.W. Ehrlich, The Lost Art of Cross-Examination 12, 25 (1970); Freed-
man, "Professional Responsibility of the Criminal Defense Lawyer: The Three Hard-
est Questions," 65 Mich. L. Rev. 1474 (1966).
69. See Lawrence M. Friedman, American Law in the 20th Century 266 (2002).
See also Mirjan Damaska, Evidence Law Adrift 94 (1997), for a broad discussion of
the many reasons why the American adversarial system of fact-finding is different
from a rational and truth-oriented inquiry about the facts.
70. See id. at 101.