ABSTRACT

Although legal reasoning is central to the dispute resolution process, very little is actually known about how judges and arbitrators approach that particular task. Instead, anecdotes and assumptions abound, with international disputes suffering the most, due to longstanding but largely unsupported theories about purported differences between common law and civil law reasoning.

This Article provides unique and useful insights into the legal reasoning process by presenting data from a recent large-scale empirical study of legal reasoning in commercial disputes. The discussion begins by considering why judges and arbitrators from common law and civil law countries write reasoned rulings before moving on to analyse and compare the importance of different types of legal authorities in substantive and procedural disputes. In so doing, the Article not only tests empirical findings against various theoretical assumptions relating to legal reasoning but also identifies the extent to which common law and civil law respondents diverge in their beliefs and practice. The analysis closes with a discussion of how providers of judicial and arbitral education can assist common law and civil law judges and arbitrators seeking to improve the way they approach legal reasoning.

I. INTRODUCTION

One of the hallmarks of international commercial arbitration involves the routine use of fully reasoned awards, a phenomenon that makes international commercial arbitration comparable to litigation in terms of complexity, sophistication and nuance.¹ However, very little information exists as to what constitutes a reasoned award in the international commercial context,² how international commercial awards differ from judicial decisions (if at all) and whether and to what extent any distinctions in legal reasoning arise across the common law-civil law divide.³

This final item is of particular interest, both as a scholarly and a practical matter, given the purposeful blend of common law and civil law elements and the predominance of both common law and civil law arbitrators in international commercial arbitration. This Article therefore seeks to provide tangible insights into longstanding questions about the nature and quality of legal reasoning in commercial disputes by testing longstanding theoretical assumptions about alleged differences between common law and civil law reasoning against recently published empirical research. In so doing, this Article not only improves individual and institutional understanding of current practices in the field but also offers tangible suggestions on how the arbitral community might move forward in light of that knowledge.

The discussion proceeds as follows. First, section II sets forth the theoretical justification for fully reasoned awards, looking at best practices in both common law and civil law jurisdictions. Those assumptions are then tested empirically, using data generated from a recent large-scale study of legal reasoning in commercial disputes. Next, section III considers the use of legal authorities in both substantive and procedural disputes, comparing the responses of common law and civil law adjudicators with respect to each type of dispute as well as across the substantive-procedural divide and again testing longstanding theoretical assumptions against recently obtained empirical evidence. The analysis then moves in section IV to a discussion of how legal reasoning among both common law- and civil law-trained adjudicators might be improved through continuing professional education. The suggestions contained in this section are based not only on best practices in judicial and arbitral education but also on data gleaned from a recent empirical study of judges and arbitrators from around the world. Finally, the Article concludes in section V with a number of forwardlooking observations.

Before beginning, it is useful to provide some basic information about the research methodology associated with the empirical data presented in this Article.⁴ Appendices I through IV contain four charts derived from a large-scale, multi-pronged study recently conducted by the author of legal reasoning in commercial disputes.⁵ The information found in the Appendices was generated from an anonymous online survey of 465 judges and arbitrators from around the world who were asked to discuss legal reasoning in national and international commercial disputes.⁶

Respondents included individuals from both the common law and civil law traditions.⁷ The common law countries represented, calculated to the hundredth of a percentage point based on the total number of respondents indicating they were qualified in a common law jurisdiction (N=322), include Australia (0.62% ), the Bahamas (0.31%), Canada (4.35%), Hong Kong (0.93%), India (0.62%), Singapore (0.31%), the United Kingdom (6.83%) and the United States (86.02%). The civil law countries represented, calculated to the hundredth of a percentage point based on the total number of respondents indicating they were qualified in a civil law jurisdiction (N=83), include Argentina (2.41%), Austria (2.41%), Belarus (2.41%), Belgium (7.23%), Bolivia (1.20%), Brazil (7.23%), Colombia (2.41%), Denmark (1.20%), Egypt (1.20%), Finland (1.20%), France (13.25%), Germany (9.64%), Greece (1.20%), Guatemala (1.20%), Italy (4.82%), Latvia (1.20%), Mexico (3.61%), the Netherlands (1.20%), Norway (1.20%), Paraguay (1.20%), Peru (1.20%), Spain (8.43%), Sweden (3.61%), Switzerland (14.46%), Turkey (1.20%), Uzbekistan (1.20%), and Venezuela (2.41%).

Three of the four charts (Appendices I-III) involve questions where survey participants were asked to indicate the relative importance of various items, where five was considered "very important" and one was considered "not important at all".⁸ The fourth chart (Appendix IV) involved a question where respondents were asked to indicate whether certain topics must be addressed (rated five out of five) or need not be addressed (rated one out of five). The charts separate common law from civil law responses and report the weighted average of responses relating to each entry, followed by the standard deviation in parenthesis, followed by the total number of responses relating to that particular item and concluding with the rank of that item within the cohort’s responses for that question.

Potential differences between common law and civil law responses were analysed by considering: (1) the relative standing of the various responses to each question within each cohort (i.e. common law respondents or civil law respondents), (2) the relative standing of the responses to each question between the two cohorts and (3) the statistical significance of the difference between intensity markings of the two cohorts with respect to individual responses. The cross-group intensity comparisons were reported in a separate column and relied on a t-statistical test, which compares the mean of two independent populations (μ1 and μ2) along a single variable to determine whether any difference exists other than that attributable to chance. Statistical significance was deemed to exist in cases with an alpha (a1) level of p<.05 or lower.⁹ The study reports p levels of .05 (designated by a single asterisk), .01 (designated by two asterisks) and .001 (designated by three asterisks).¹⁰

While the study provides a number of very interesting insights into common law and civil law perspectives on legal reasoning, it does not focus solely on disputes arising in international commercial arbitration. As noted previously, the study included both judges and arbitrators, and allowed participants to discuss commercial disputes that were both domestic and international in nature.¹¹ Although the number of respondents who identified solely as judges was dwarfed by the number of individuals who had acted as arbitrators, either exclusively or in addition to acting as a judge, the text will refer to both judicial decisions and arbitral awards (jointly characterised as "reasoned rulings") except when the two can and should be distinguished.¹² Notably, the inclusion of judges and domestic arbitrators in this study is by no means problematic to the current analysis, since the goal of the Article is to identify whether and to what extent any differences in legal reasoning arise across the common law-civil law divide. Indeed, inclusion of judges and domestic arbitrators may give a more realistic view of standard practices in common law and civil law jurisdictions, since those individuals are less likely to reflect the type of "transnationalised" mindset that some commentators believe exists in international commercial arbitration.¹³

II. REASONED RULINGS: DEFINITIONS, DISTINCTIONS AND MOTIVATIONS

II.A. Theoretical Considerations

II.A.1. Defining reasoned rulings

The first matter to consider involves the question of what constitutes a fully reasoned judicial decision or arbitral award in international commercial disputes. In cross-border litigation, the matter is determined by national practice, as outlined in the relevant rules of civil and judicial procedure, while in arbitration, the matter is typically defined by institutional rules.

However, in both settings, the governing provisions are often quite vague and simply indicate that an award should include "reasons", with little if any explanation as to what is entailed by that term.¹⁴

To some extent, the absence of detail regarding the shape and content of a reasoned ruling may be the result of the difficulties inherent in describing what constitutes adequate legal reasoning as an abstract proposition. Indeed, it is often easier to identify specific examples of fully reasoned decisions and awards than it is to provide a categorical definition of what constitutes adequate legal reasoning.¹⁵

Some assistance may be gained from the world of investment arbitration, where Tai-Heng Cheng and Robert Trisotto have argued that the reasoning requirement reflected in institutional rule sets

is in fact a reasoning standard. Disagreements among [annulment] committees about whether the standard should be high or low are… fundamentally about what methods of reasoning are acceptable. The high standard countenances only reasoning that is correct on the law and facts and the rational derivation of outcomes therefrom; the low standard tolerates reasoning that is incorrect due to mistakes in the law or facts, so long as the reasoning is internally consistent; and the intermediate standard requires coherence and permits errors of law and fact, so long as these errors are reasonable errors.¹⁶

Similar types of studies have not yet been conducted in international commercial arbitration, likely because of the difficulties associated with obtaining original awards in their entirety. However, Gary Born has noted that being correct as a matter of law (i.e. the high standard identified by Cheng and Trisotto in the investment context) is "not a requirement for a well-reasoned [commercial] award: bad or unpersuasive reasons are still reasons, and satisfy statutory requirements for reasoned awards."¹⁷

Additional guidance may be obtained from domestic sources. For example, the State of Pennsylvania has described a reasoned ruling in an administrative law dispute in the following terms:

[A]ll findings of fact shall be based upon sufficient competent evidence to justify same. All parties to an adjudicatory proceeding are entitled to a reasoned decision containing findings of fact and conclusions of law based upon the evidence as a whole which clearly and concisely states and explains the rationale for the decisions so that all can determine why and how a particular result was reached. The… judge shall specify the evidence upon which the… judge relies and state the reasons for accepting it in conformity with this section. When faced with conflicting evidence, the… judge must adequately explain the reasons for rejecting or discrediting competent evidence. Uncontroverted evidence may not be rejected for no reason or for an irrational reason; the… judge must identify that evidence and explain adequately the reasons for its rejection.¹⁸

As useful as this definition is, it may still fall somewhat short in practice, since finding "the appropriate methodology for distinguishing questions of fact from questions of law [is], to say the least, elusive."¹⁹ Indeed, "the practical truth [is] that the decision to label an issue a ‘question of law’, a ‘question of fact’, or a ‘mixed question of law and fact’ is sometimes as much a matter of allocation as it is of analysis."²⁰

Although reasoned rulings have long played a central role in the common law legal tradition, civil law jurisdictions also consider fully reasoned legal decisions to be essential to procedural justice.²¹ However, a reasoned decision in a civil law nation can not only look very different from a reasoned decision in a common law country, it can also vary significantly in comparison to other civil law decisions.²² At one end of the spectrum lie French judgments, which are usually "formulated in a single sentence, including several ‘whereas-es’ (attendus)".²³ Although some common law commentators appear to believe that "French practice belies the notion that well-reasoned [apparently meaning fully reasoned] opinions are in some sense necessary",²⁴ French decisions have been described as being "consistent with—and probably produced by—the primacy of text, conceptualism, and deduction, as well as the post-revolutionary caution on the part of judges not to exceed their limited powers."²⁵

As well-known as the French style may be, it is by no means standard within the civil law legal tradition. For example, it has been said that

[w]hoever compares the arguments of a decision of a German Landgericht with those of a Dutch rechtbank will be impressed by the length and thoroughness of the German argument on the one hand, the straightforward, paper-saving decision of the Dutch court on the other. In appeal courts and before the highest courts the differences in elaborateness are even more apparent. German legal style is much more differentiated, scholarly worded; the style of Dutch courts is pragmatic....²⁶

Furthermore, the French approach does not appear to have been routinely adopted in international commercial arbitration, despite the esteem with which France is held in the field.²⁷ Instead, reasoned awards in international commercial arbitration more frequently resemble judicial decisions generated by courts in common law countries and in civil law jurisdictions like Germany.²⁸

The difficulties associated with defining reasoned rulings both within and between different legal systems suggest that further analysis should proceed on a functional basis.²⁹ This approach is particularly useful because it not only overcomes differences between litigation and arbitration but also between common law and civil law traditions.³⁰ The discussion will therefore move to considering structural and non-structural rationales for reasoned rulings.

II.A.2. Structural rationales for reasoned rulings

Perhaps the most well-known structural rationale for reasoned rulings involves the common law concept of stare decisis, which requires judges to adhere to legal principles established in judgments rendered by courts whose decisions are binding upon the ruling court.³¹ Because stare decisis does not technically apply in civil law jurisdictions or in international commercial arbitration, this rationale may not initially appear relevant outside common law courts.³² However, judges in a number of civil law countries routinely follow the decisions of higher level courts, even if the principle of precedent is not part of that particular legal system, so as to promote predictability and consistency in the law.³³ Arbitral awards can also generate a type of persuasive authority, particularly matters involving arbitral procedure, again as a result of the need for predictability and consistency.³⁴ Thus, reasoned awards and civil law decisions may be said to meet this first type of structural purpose, even if stare decisis does not technically apply outside of common law courts.

Reasoned rulings serve other structural purposes. For example, courts in both common law and civil law countries use reasoned rulings to give context to lower court decisions and thereby help appellate courts determine whether and to what extent the judgment below should be upheld.³⁵ This rationale might also initially appear inapplicable to international commercial arbitration, since most jurisdictions do not allow courts to review the merits of an arbitral award.³⁶ However, some jurisdictions, most notably England, allow judicial appeals of international awards, which could be seen as providing arbitrators with a strong incentive to render fully reasoned awards in arbitrations seated in England.³⁷

Although arbitral awards are seldom reviewed on the merits, they may be subject to other types of post-award scrutiny. The most likely of these types of actions involve efforts to vacate or enforce an arbitral award, domestically or internationally.³⁸ Although these actions typically focus on procedural concerns, the likelihood of a challenge being brought in the first place may be affected by the quality of the reasoning found in the underlying award.³⁹ For example, a well-written, fully reasoned award may persuade the losing party that a decision is well-supported, even if the outcome is negative.⁴⁰ Alternatively, a fully reasoned award may diminish the likelihood of a judicial challenge by eliminating certain grounds for non-enforcement.⁴¹

Another possible action involves collateral or parallel proceedings.⁴² Although the law concerning preclusion and collateral estoppel are not as well developed in arbitration as in litigation,⁴³ a court or arbitral tribunal may find itself unable to give preclusive effect to an award that is not fully reasoned, since such an award may not indicate whether a particular issue was fully and fairly addressed in the earlier proceeding.⁴⁴

The final type of post-award procedure to consider involves arbitral appeals. Over the last few years, several arbitral organisations have adopted special rule sets governing appellate arbitration.⁴⁵ Arbitral appeals have important ramifications for the award writing process, both at first instance and on appeal.⁴⁶ For example, arbitrators hearing a dispute as an initial matter may need to be increasingly aware of the quality of their awards both to avoid creating an appealable issue⁴⁷ and to provide an appellate tribunal with a solid understanding of how and why the initial decision was made.⁴⁸

II.A.3. Non-structural rationales for reasoned rulings

Although lawyers are generally familiar with structural rationales for reasoned rulings, various non-structural rationales also exist. First, reasoned rulings have been said to constitute part of a "procedural trinity that is necessary to establish the rule of law", although the strength and nature of this element may vary somewhat across the common law-civil law divide.⁴⁹ For example, Mathilde Cohen has observed that,

[i]n the common law tradition, judicial reason-giving does not attach to individual right holders. Litigants or defendants have no enforceable right to reasoned decisions; a judge’s failure to give reasons does not amount to reversible error. The validity of the judgment is evaluated independently from the reasons given for it. Unreasoned decisions are protected by the "harmless error" doctrine. The rationale behind this rule appears to be judicial economy. No appeal lies on the grounds that a court gave inadequate or insufficient reasons. A successful appeal must contend that the judgment was incorrect and thus warrants reversal.⁵⁰

In contrast, some civil law jurisdictions consider the absence of reasons to be "a procedural defect that warrants voiding the judgment altogether."⁵¹

Arbitration also respects and protects procedural fairness, with Catherine Rogers recognising that "the product of international arbitral decisionmaking is justice".⁵² While some jurisdictions allow parties in arbitration to waive the right to a reasoned award, other countries do not.⁵³ Furthermore, even when a waiver is allowed, parties in international commercial disputes seldom take advantage of that opportunity.⁵⁴ Thus, while the reasoning requirement in international commercial arbitration may differ from similar standards applicable in litigation,⁵⁵ it is clear that arbitrators must provide some minimal level of reasoning once the parties have requested a reasoned award.⁵⁶ As it turns out, the length and details commonly associated with reasoned awards in international commercial arbitration suggest that many arbitral tribunals are far exceeding any minimum requirements.⁵⁷

Second, reasoned rulings are believed to improve the quality of the decisionmaking process and consequently of the decision or award itself.⁵⁸ As U.S. Circuit Judge Richard Posner once said, "[r]easoning that seemed sound when ‘in the head’ may seem half-baked when written down, especially since the written form of an argument encourages some degree of critical detachment in the writer, who in reading what he [or she] has written will be wondering how an audience would react."⁵⁹ By forcing judges and arbitrators to articulate their thought processes, reasoned rulings help "rationalize the… process", "safeguard against arbitrary decisions", "prevent consideration of improper and irrelevant factors", "minimize the risk of reliance upon inaccurate information", and "attain[…] institutional objective[s] of dispensing equal and impartial justice" while simultaneously "demonstrat[ing] to society that these goals are being met".⁶⁰

Third, reasoned rulings can enhance the legitimacy of the dispute resolution process, either in general or with respect to the immediate proceeding, by demonstrating the integrity of the judicial or arbitral endeavour.⁶¹ This factor may be of particular relevance to international arbitration, which has come under increasing attack in recent years, with critics indicating concerns about the lack of transparency and the supposedly preferential treatment of large, multinational firms.⁶² Fully reasoned arbitral awards can help assuage fears about the quality of justice provided by arbitrators by detailing the high degree of care and knowledge that go into the decision-making process.

Fourth and finally, reasoned awards provide parties with a more robust explanation of why the arbitral tribunal decided as it did.⁶³ This feature can be quite important, since empirical studies have shown that "the perceived fairness of arbitration hearings significantly predicts litigant decisions to accept an arbitration decision",⁶⁴ which suggests that parties who believe that they have not been fully "heard" during the arbitration, as reflected by the quality or content of the reasoning enunciated in the written award, might be more inclined to mount a challenge to that award.⁶⁵

II.B. Empirical Findings

Having set forth the theoretical reasons for reasoned awards, the next step involves testing those assumptions against empirical observations. Appendix I contains responses to a survey question asking judges and arbitrators from both common law and civil law countries to identify factors used to determine how detailed a reasoned decision or award should be.⁶⁶ The question was framed in this way to avoid confusion among the respondents about whether a reasoned ruling should be written in the first place, since that matter is often controlled by external factors, particularly in arbitration, where arbitration agreements or arbitral rules frequently mandate the type of award that is to be rendered.⁶⁷ While judges may be able to exercise slightly more discretion as to whether they will write a fully reasoned decision, external influences may nevertheless exist in the form of pressure from chief judges and/or the court’s internal norms of judicial procedure.⁶⁸

The question focused on nine different alternatives drawn from the theoretical literature on legal reasoning and asked respondents to indicate how important each item was, based on a five-point scale, where one was the least important and five was the most important. The resulting data was analysed with respect to (1) the relative standing of the nine options within each cohort (i.e. common law respondents or civil law respondents), (2) the relative standing of the nine options between the two cohorts and (3) the statistical significance of the difference between intensity markings of the two cohorts.

As Appendix I shows, the responses to this question were remarkably similar across the common law-civil law divide. Not only were the first five responses identical as between common law and civil law participants, but the intensity of the first five responses showed no statistically significant difference between the two groups. This information is useful in offsetting the common assumption that structural considerations (such as those relating to the need to create and establish common law precedent) are the most important motivations for reasoned rulings.

This conclusion is further supported by the nature of the top five responses. For example, the most highly ranked responses focused on content-based considerations, with the number-one answer citing the difficulty, complexity or novelty of the dispute as a matter of fact (rather than of law) and the number-two answer, by a very small margin of intensity, citing the difficult, complexity or novelty of the dispute as a matter of law (rather than of fact). While this information does not address the question of whether respondents engaged in deductive versus inductive reasoning, a concept that has long been believed to differentiate common law versus civil law reasoning,⁶⁹ the data nevertheless suggest that judges and lawyers approach the writing process similarly across the common law-civil law divide.

The third most common response in both groups indicated that the act of writing helped the respondent think through the materials, thereby increasing the likelihood of reaching the right outcome. While common law lawyers rated this item slightly higher in absolute terms, giving it a 4.21 out of 5.00 as opposed to civil law lawyers’ 3.90, the difference was not statistically significant. What is intriguing, however, is the acknowledgement that the act of writing helps the analytical process, something that had been posited as a theoretical matter.⁷⁰

The fourth most common response indicated that judges and arbitrators believe it is important to make sure that the losing party feels heard, something that has also been said to be important by theoretical commentators.⁷¹ Interestingly, survey respondents also indicated that it was important to make sure that the winning party felt heard but at a much lower level of intensity, with that option appearing as the eighth highest-ranking alternative out of nine among both common law and civil law respondents.

The fifth most important factor involved the desire to increase the perceived legitimacy of the particular award at issue, a concept that is related to commentators’ observations about the connection between reasoned rulings and legitimacy.⁷² Although the intensity of this choice varied slightly between the two groups, earning a 3.92 out of 5.00 from common law lawyers and only a 3.67 from civil law lawyers, the difference was not statistically significant.

At this point, the hierarchical rankings began to diverge. Only two demonstrated statistically significant differences (p<.05) between common law and civil law respondents. One involved concerns about whether the losing party would challenge the resulting decision or award, with civil law lawyers giving this item intensity marks of 3.60 out of 5.00 and ranking it as the sixth most important factor to their determination of how detailed a reasoned ruling should be, while common law lawyers gave this factor a mark of 3.03 and ranked it as the least most important factor (ninth out of nine) to the question of how detailed a reasoned ruling should be. The other area of divergence involved the desire to increase the perceived legitimacy of the dispute resolution process in question (i.e. arbitration or litigation as a general proposition), as opposed to the legitimacy of the particular decision or award at issue. Common law lawyers ranked this as the seventh most important factor to the question of how detailed a reasoned ruling should be and gave it an intensity marking of 3.34 out of 5.00, while civil law lawyers ranked this factor as ninth out of nine, with an intensity ranking of 2.90.

Interestingly, both common law and civil law lawyers considered the nature of the legal dispute in question (i.e. the type of substantive claims at issue) to be relatively unimportant to the determination of how detailed a reasoned decision or award should be. Common law lawyers ranked this option as their sixth most important factor, with an intensity mark of 3.68 out of 5.00, and civil law lawyers ranked it as their seventh most important factor, with an intensity mark of 3.57. However, the difference in intensity ratings was not statistically significant.

The data generated by this question is useful in that it confirms a number of theoretical assumptions about the motivations behind reasoned rulings and demonstrates a high degree of consistency across the common law-civil law divide. However, it is possible to gain even more information about the nature of the reasoning process itself, as discussed in the next subsection.

III. USE OF LEGAL AUTHORITIES IN SUBSTANTIVE AND PROCEDURAL DISPUTES

III.A. Theoretical Considerations

The next item to consider involves the use of legal authorities by common law- and civil law-trained individuals when resolving substantive and procedural disputes. Although a number of assumptions exist about the differences between common law and civil law reasoning,⁷³ one of the most frequently mentioned involves claims about the types of legal authority that are considered most persuasive in particular jurisdictions. For example, conventional wisdom suggests that common law lawyers find judicial decisions to be the most important type of legal authority while civil law lawyers are said to downplay the importance of such materials.⁷⁴ Conversely, commentators frequently claim that civil law lawyers have a heightened regard for scholarly commentary as compared to the relatively minimal role that scholarship plays in common law jurisdictions.⁷⁵

Although most comparative analyses relating to the use and priority of legal authorities in common law and civil law jurisdictions focus on questions of substantive law,⁷⁶ some comparative procedural analyses exist courtesy of international commercial arbitration.⁷⁷ These latter studies distinguish between seven types of legal authority: (1) international conventions and treaties; (2) national arbitration laws; (3) arbitral rules, such as those published by arbitral institutions, and other forms of arbitral soft law; (4) the law of the dispute, which can include agreements between the parties, procedural orders issued by the arbitral tribunal and previously rendered arbitral awards involving the parties; (5) arbitral awards not involving the parties to the immediate dispute, meaning awards that have been published in whole or in part in arbitral reporters or as made public as exhibits to judicial proceedings; (6) judicial decisions; and (7) scholarly commentary.⁷⁸ This list is useful because it recognises the importance of both public and private sources of law and therefore formed the basis for survey questions regarding the use of legal authorities in substantive and procedural disputes.

III.B. Empirical Findings

Over the last few decades, questions have arisen about longstanding theoretical assumptions about the relative weight of different legal authorities in common law and civil law legal systems.⁷⁹ Not only have some commentators suggested that statutes are now nearly as important in common law countries as they are in civil law countries, but some civil law jurisdictions are said to rely on judicial decisions in much the same way that common law jurisdictions do.⁸⁰ As intriguing as these hypotheses are, they have not been supported by rigorous empirical research.⁸¹ However, that situation has changed as a result of recent studies on legal reasoning in commercial disputes.⁸² Not only does this new empirical work consider potential differences between the use of legal authorities in common law and civil law countries in both substantive and procedural disputes, it also evaluates potential differences across the substantive-procedural divide.

The data is reproduced in Appendices II and III. The analysis begins with Appendix II, which considers the importance of ten different types of legal authorities associated with the determination of substantive disputes. In some ways, the survey findings confirm the standard theoretical understanding of how legal authorities are used and weighed in common law and civil law jurisdictions. For example, the five most important authorities for common law judges and arbitrators were, in decreasing order of importance: (1) judicial decisions from the jurisdiction whose law controls the issue; (2) domestic regulations, statutes or administrative rules; (3) judicial decisions from jurisdictions other than the jurisdiction whose law controls the matter; (4) legislative histories of domestic regulations, statutes or administrative rules; and (5) scholarly works. However, judges and arbitrators from the civil law tradition indicated that the five most important authorities for determining substantive disputes were, in decreasing order of importance: (1) domestic regulations, statutes or administrative rules; (2) international treaties; (3) judicial decisions from the jurisdiction whose law controls the matter; (4) scholarly works; and (5) soft law. While the relatively high status of international treaties could be attributed to the fact that a relatively high proportion of civil law respondents indicated that they were focusing on international rather than domestic disputes,⁸³ it is also consistent with the fact that most civil law nations adopt a monist perspective to international law.⁸⁴

As useful as the hierarchical data is in comparing how common law and civil law lawyers conduct legal reasoning in substantive disputes, the intensity marks are equally illuminating. Seven of the ten types of legal authority showed a statistically significant difference between the two cohorts. Five items reflected differences at the highest level of statistical significance (p<.001): international treaties, which featured an intensity mark of 2.62 out of 5.00 from common law respondents and an intensity mark of 4.37 from civil law respondents; judicial decisions from the jurisdiction whose law controls the matter, which featured intensity marks of 4.76 from common law respondents and 4.35 from civil law respondents; judicial decisions from jurisdictions other than the jurisdiction whose law controls the matter, which were given intensity marks of 3.19 by common law respondents and 2.39 by civil law respondents; legislative history of international treaties (travaux preparatoires), which earned intensity marks of 2.00 from common law respondents and 3.00 from civil law respondents; and scholarly works, which won intensity marks of 2.91 from common law respondents and 3.54 from civil law respondents. Two type of authorities—arbitral awards involving parties other than those involved in the instant dispute and soft law—reflected a statistically significant degree of difference between common law and civil law respondents at an intermediate level of confidence (p<.01).⁸⁵ These results, like those generated from the hierarchical rankings, confirm the standard understanding of how common law and civil law training affects legal reasoning in substantive disputes.

However, the inquiry did not stop there. Instead, the survey also asked judges and arbitrators to identify the importance of different types of legal authorities when deciding procedural disputes, offering the same list of ten authorities that were discussed with respect to substantive disputes. Interestingly, survey respondents did not place the authorities in the same hierarchical order that was seen with substantive disputes, nor did the respondents give the individual alternatives the same intensity ratings.

As Appendix III shows, common law and civil law respondents showed a much higher level of convergence with respect to the relative importance of procedural authorities as compared to substantive authorities. First, the two cohorts agreed on the two most important types of procedural authority: (1) judicial decisions from the jurisdiction whose law controls the matter and (2) domestic regulations, states or administrative rules which, in the procedural context, would mean rules of civil procedure (in litigation) and national laws on arbitration (in arbitration). Second, the two groups came to very similar conclusions about the importance of scholarly works, with common law respondents putting that particular authority in a tie for third place along with arbitral awards involving the parties to the instant dispute and civil law respondents placing scholarly works in fourth position. The remaining authorities reflected varying levels of divergence between the different types of authority.

Useful information was also generated with respect to intensity rankings. For example, the intensity marks for the first two categories of legal authorities (i.e. those on which the respondents were in agreement as to relative importance) did not differ to a statistically significant degree, even though six of the remaining eight items did. Four of those additional authorities—international treaties, legislative history of international treaties, scholarly works and soft law—differed at the highest level of confidence (p<.001), while two authorities—arbitral awards involving parties other than those involved in the immediate dispute and judicial decisions from jurisdictions other than the jurisdiction whose law controls the matter— differed at an intermediate level of confidence (p<.01).

As useful as it is to analyse Appendix II and Appendix III independently, comparing the two charts yields a number of unique insights not reflected in any other known studies. First, the data suggests that traditional views about how common law- and civil law-trained lawyers view legal authorities may be incorrect when it comes to procedural authorities. For example, respondents from the civil law rated judicial decisions more highly in procedural disputes than in substantive disputes, thereby resembling common law lawyers, while respondents from the common law rated scholarly works more highly in procedural disputes than in substantive disputes, thereby resembling civil law lawyers. While it remains unclear at this point why judges and arbitrators approach substantive and procedural disputes differently, this finding opens the door to additional research.⁸⁶

Second, those from civil law countries appear to be more consistent than those from common law countries with respect to their approach to legal authorities across the substantive-procedural divide. For example, civil law respondents placed the same authorities in the top five positions in both types of disputes, with only a relatively minor rearrangement of individual items in terms of relative importance.⁸⁷ In contrast, common law respondents displayed a much greater degree of variation across the substantive-procedural divide in terms of how they rated their top five authorities.

Finally, Appendices II and III show that, for the most part, civil law respondents tended to rate substantive authorities at a slightly higher intensity rate than was the case with procedural authorities, although that difference may not be statistically significant.⁸⁸ The same phenomenon also appeared to be true with respect to the common law respondents, although such comparisons are more difficult since, as noted, common law lawyers perceived the relative importance of individual authorities in more varied terms as between substantive and procedural disputes.

As with most forms of empirical research, this data merely reports what judges and arbitrators believe with respect to the importance of various authorities. It does not confirm whether the reported beliefs are also reflected in action,⁸⁹ nor does it seek to establish why respondents believe what they do. However, the information gained from this study provides interesting avenues for future research on this issue.

IV. CONTINUING PROFESSIONAL EDUCATION ON LEGAL REASONING

IV.A. Theoretical Considerations

Although relatively few studies exist regarding the need for and nature of continuing professional education of judges and arbitrators, recent empirical research indicates that this is a field that deserves more attention. Indeed, according to the survey discussed elsewhere in this Article, 79 per cent of common law respondents and 68 per cent of civil law respondents believed that new and experienced judges and arbitrators would benefit from additional educational programming on writing reasoned decisions and awards.⁹⁰ This data is supported by scholarly and practical commentary suggesting that the skills associated with serving as a judge or arbitrator differ significantly from those associated with acting as an advocate, with new adjudicators often finding "the ‘move from advocacy to decision, from marshalling and presenting evidence to fact-finding and synthesizing’, to be extremely challenging."⁹¹ As a result, it is "inaccurate to claim, as some authorities have, that international arbitrators can gain the necessary skillset simply through ‘observation, exposure, participation and experience’."⁹²

Although continuing education of judges and arbitrators appears to be considered increasingly important, questions exist as to what precisely should be taught. In the past, educational providers often followed the law school model of legal education, emphasizing content-based programming that discussed recent developments in substantive or procedural law or policy.⁹³ More recently, various for-profit and advocacy-oriented groups have offered judicial education courses on complex and often controversial subjects, raising concerns about the objectivity of the programming.⁹⁴ However, national and international leaders in judicial studies have begun to shift professional programming away from current events and private provider models and toward multidisciplinary curricula that include more diverse coursework, including what has been called "judge-craft".⁹⁵ This approach has been emulated to some degree in arbitral settings, with some providers offering instruction on how to write a reasoned decision or award, although most of those programmes appear to focus on relatively basic or mechanistic concerns.⁹⁶

IV.B. Empirical Findings

As judicial and arbitral education becomes more sophisticated, it becomes increasingly important to understand what judges and arbitrators need and want in order to improve their performance as adjudicators. Appendix IV shows responses to a question asking judges and arbitrators to rate eleven different types of issues that might be included in continuing professional programming on a scale of one to five, where one indicated items that did not need to be addressed and five indicated items that must be addressed.

Interestingly, very few similarities existed across the common law-civil law divide. Indeed, the only item that was ranked the same for the two cohorts was the desire for programming on how to structure reasoned decisions or awards, which was the first choice for both groups.⁹⁷ However, a statistically significant degree of intensity existed (p<.001) with respect to this particular topic, with civil law respondents being more strongly in favour of programming on this particular point, rating it at 4.64 out of 5.00, as compared to the 4.25 mark from common law respondents. This difference may be due at least in part to the fact that a relative low number (19 per cent) of civil law-trained judges and arbitrators participating in the study in question had attended a seminar on writing reasoned decisions or awards in comparison to common law-trained judges and arbitrators (57 per cent), a difference that was statistically significant (p<.001).⁹⁸

Immediately after this point, the data began to diverge. Civil law respondents rated the need to learn how to weigh different pieces of evidence as the second most important element, with common law respondents ranking that topic as only seventh out of eleven. The difference in intensity of interest in this subject was also statistically significant (p<.001), with civil law respondents giving this subject a mark of 4.08 out of 5.00 and common law respondents giving it only a 3.20.

The second most important topic for judges and arbitrators from common law countries involved how to apply the facts to the governing legal standard. This item was only fourth in importance to civil law respondents, although civil law respondents indicated a higher intensity of interest than common law respondents did, giving it a mark of 4.00 out of 5.00, as compared to 3.83 for common law respondents. However, the difference in intensity rankings was not statistically significant.

The third most highly rated topic for common law respondents—how other judges or arbitrators approach reasoned decisions or awards—was the tenth most important item for civil law respondents, although the difference in intensity ranking was not statistically important. The third most important item for civil law respondents—how to identify the governing legal standard—was the fifth most important factor to common law respondents, and the intensity ranking (4.03 out of 5.00 for civil law respondents and 3.39 for common law respondents) reflected a statistically significant difference (p<.01).

Notably, none of the respondents indicated that they wanted more information involving current developments in law or practice.⁹⁹ This suggests that providers of judicial and arbitral education would do well to shift their focus away from law-school type programming, where the content of the law is of primary importance, and toward a more skills-based curriculum consistent with contemporary best practices in the field of judicial studies.

V. CONCLUSION

International commercial arbitration has long been praised for transforming the once-arcane field of comparative law into "a kind of ‘living comparison’ of laws… emerging from the continuous communication between persons educated in different intellectual and legal contexts."¹⁰⁰ Indeed, international commercial arbitration has become one of the major "clients" of comparative law as arbitrators and advocates seek to improve their understanding of how foreign legal systems operate.¹⁰¹ Unfortunately, most of the knowledge about the differences between common law and civil law legal systems appears to be somewhat outdated and based more on anecdote rather than evidence.

This Article has sought to counteract conventional wisdom by testing various longstanding assumptions about legal reasoning in common law and civil law countries against recent empirical research in that field. While much remains to be done,¹⁰² this Article has nevertheless provided new and important insights into comparative commercial dispute resolution. In so doing, this Article hopes not only to provide parties and practitioners with useful data that will assist them in presenting their cases to judges and arbitrators, but also to help scholars develop new avenues of theoretical and empirical analysis.

APPENDIX I¹⁰³

Factors important to the determination of how detailed a reasoned decision or award should be (weighted averages, followed by standard deviation in parenthesis, sample (N) size and ranking within the relevant cohort) Statistical significance is indicated by asterisk, where one asterisk (*) indicates p<.05, two asterisks (**) indicate p<.01, and three asterisks (***) indicate p<.001

APPENDIX II¹⁰⁴

Importance of different types of legal authorities when deciding substantive disputes (weighted averages, followed by standard deviation in parenthesis, sample (N) size and ranking within the relevant cohort)

Statistical significance is indicated by asterisk, where one asterisk (*) indicates p<.05, two asterisks (**) indicate p<.01, and three asterisks (***) indicate p<.001.

APPENDIX III¹⁰⁵

Importance of different types of legal authorities when deciding procedural disputes (weighted averages, followed by standard deviation in parenthesis, sample (N) size and ranking within the relevant cohort)

Statistical significance is indicated by asterisk, where one asterisk (*) indicates p<.05, two asterisks (**) indicate p<.01, and three asterisks (***) indicate p<.001.

APPENDIX IV¹⁰⁶

Issues to be included in educational programming on reasoned decisions and awards (weighted averages, followed by standard deviation in parenthesis, sample (N) size and ranking within the relevant cohort)

Statistical significance is indicated by asterisk, where one asterisk (*) indicates p<.05, two asterisks (**) indicate p<.01, and three asterisks (***) indicate p<.001

---

---