I. Introduction

A. Background

1. At the October 2015 meeting of the ICC Commission on Arbitration and ADR, Toby Landau, QC delivered a guest speech entitled Unreliable Recollections, False Memories and Witness Testimony. In that speech, Toby Landau gave the Commission an insight into the science of human memory, highlighting its fragile and malleable nature and the ease with which memories can become unwittingly corrupted. He then questioned whether the practices that are commonly adopted for the preparation and presentation of fact witness evidence in international arbitration are themselves potentially corrupting the very evidence that arbitral tribunals rely upon for the fair resolution of disputes.
2. Toby Landau’s speech was not just entertaining and thought-provoking. It invited further enquiries to be made. If indeed these risks exist, what practical steps might be taken to improve the probative value of witness evidence in international arbitration?
3. The Commission decided to take up this invitation and created the Task Force Maximizing the Probative Value of Witness Evidence. Its mandate was to look at the science (with input from eminent psychologists specialising in human memory), at arbitral practice (with input from Task Force members specialising in international arbitration from around the globe) and to consider whether modifications could be made to current practices, or alternative approaches could be adopted, in order to enhance the probative value of fact witness evidence in international arbitration, particularly as it is affected by memory.

B. Summary of conclusions

4. This Report describes the work undertaken by the Task Force, including the independent study that it commissioned by Dr Kimberley A. Wade of the Department of Psychology at the University of Warwick in England. It considers the pertinence of memory issues in international arbitration and sets forth some measures that can be taken to protect witnesses’ memories.
5. The Report sets out the Task Force’s conclusions and recommendations (in Section VI), which can briefly be summarised as follows:
1. Science shows that the memory of an honest witness who gives evidence in international arbitration proceedings can easily become distorted and may therefore be less reliable than the witness, counsel or the tribunal expects. Greater awareness of the circumstances in which memory distortion is likely to occur and the measures that can be taken to avoid such distortion will be a key step forward for all participants in the arbitration process.
2. There are many steps (as set out in detail in Section V) that can be taken by witnesses, in-house counsel, outside counsel and arbitral tribunals to reduce the risk of distortions of witness memory and to better assess the weight to be given to witness evidence in the light of any distortions. In some cases, however, those steps will be impractical, or may actually reduce rather than enhance the accuracy of the evidence and the efficiency of its presentation. A case by case (and potentially witness by witness) assessment is therefore required to determine which steps are appropriate.
3. Witness evidence is presented in arbitration proceedings for different purposes, many of which do not rely upon the accuracy of witness memory (as discussed in Section IV). Where the accuracy of witness memory is not relevant, neither are concerns regarding memory corruption.

C. Why is this Report relevant to arbitration practitioners?

6. The Commission considered the work of this Task Force to be a valuable contribution to the international arbitration community because witness evidence is not only a common feature in international arbitration, but it often takes a central role in arbitration proceedings:
1. The preparation of witness statements is a significant task that consumes the time and attention of both the witnesses involved and the legal counsel representing the party relying on the evidence.
2. The preparation of witness statements can be costly, particularly where the statements are in narrative form and give an extensive account of the factual background to the dispute.
3. The presentation of oral evidence is one of the main functions of a ‘final’ hearing. It tends to consume the majority of the time taken at such a hearing, often amounting to several days.
4. The costs associated with the presentation of oral evidence are usually significant. The witnesses must take time away from their usual occupations to attend the hearing and often incur out of pocket travel and accommodation costs. They may also spend time preparing to give their evidence. The lawyers spend time preparing the witnesses called by their client and the cross-examination of the other side’s witnesses. The tribunal will read all the witnesses' statements in advance of the hearing so that it can understand the evidence given in cross-examination and ask follow-up and questions to seek clarification as appropriate.
5. The decision of the tribunal on the merits of the case will often turn, in varying degrees, on the witness evidence that has been presented. One of the important judgments that tribunal members will often have to make is to determine the credibility of a witness and the weight to be given to a witness’ evidence.

7. The time, effort and cost that are put into the collation and presentation of witness evidence are considered to be justified on the basis that they assist the tribunal reach a just decision in the case. That justification is called into question if the witness evidence is not as reliable as the participants in the process (and most importantly the tribunal) believe it to be.
8. Historically, different legal cultures have taken different approaches to the presentation of witness evidence in both litigation and arbitration proceedings. In broad terms, civil law jurisdictions have placed less focus on witness evidence and more focus on documentary evidence. Some civil law jurisdictions are so sceptical regarding the value of witnesses who are employed by a party to the proceedings that evidence from such witnesses is not admissible at all. In contrast, common law jurisdictions have placed significant focus on witness evidence and it is from those jurisdictions that the traditions of narrative witness statements and the cross-examination of witnesses have emerged.
9. As with other areas of arbitral procedure, the international arbitration community has sought to harmonise its approach to the presentation of witness evidence so as to provide what some refer to as a ‘level playing field’. The IBA Arbitration Committee in particular has done significant work in this regard, most notably in publishing the ‘IBA Rules on the Taking of Evidence in International Arbitration’ and the ‘IBA Guidelines on Party Representation in International Arbitration’. The latter of these two documents states:

‘Many international arbitration practitioners desire more transparent and predictable standards of conduct with respect to relations with Witnesses and Experts in order to promote the principle of equal treatment among Parties.’¹

10. This harmonisation has resulted in a well-accepted default approach to the presentation of witness evidence in international arbitration proceedings which includes the preparation and exchange (simultaneous or sequential) of narrative statements from each witness of fact.
11. The ‘IBA Guidelines on Party Representation in International Arbitration’² provide that a party representative may:
• assist witnesses in the preparation of their witness statements, whilst seeking to ensure that a witness statement reflects the witness’ own account of relevant facts, events and circumstances; and
• meet or interact with witnesses in order to discuss and prepare their prospective testimony.
12. These practices are already widely used, which was no doubt a factor in the IBA Arbitration Committee’s decision to propose them as accepted ‘best practice’.
13. The international arbitration community has therefore developed a well-accepted practice for the presentation of witness evidence which has undoubtedly assisted in levelling the playing field. However, what has not been done –at least in the field of international arbitration – is to consider whether the best practices that we have developed produce witness evidence that has the probative value expected of it and justifies the effort and expense that go into its preparation.

D. Summary and structure of the Report

14. Scientific studies and human experience show that witness memory is imperfect and is subject to possibly distorting influences as soon as it is formed. Memory is malleable - it is not akin to a fixed image that is 'retrieved' when needed, but rather it is a dynamic process that can be affected by subsequent events.
15. As the science establishes that witness memory can become altered or contaminated, the Task Force examined ways to reduce the distortions, and in particular, ways in which the participants in the arbitral process (witnesses, in-house and outside counsel, arbitrators) can reduce the distortions in the memory process.
16. For instance, practices that are not uncommon in modern arbitral practice, such as interviewing witnesses in a group rather than individually, can distort a memory. Also, lawyers arrive ‘on the scene’ at different points after the memory is formed, and one of the aims of this Report is to reduce the risk of memory distortion as early as possible in the process. The suggestions include steps that can be taken to preserve memory such as the maintenance of contemporaneous documents (even before the lawyers arrive on the scene) as well as practices designed to reduce the distortion-causing effects of lawyer/witness interaction and other practices such as collective witness preparation.
17. When addressing how to enhance the value of witness evidence, one must be careful to distinguish between inaccuracies or distortions that result from the imperfections of the memory process and deliberate efforts by a witness to distort or shade the truth. This Report identifies measures that can help to reduce or prevent the inaccuracies or distortions that result from the former, namely, unintentional but nevertheless real, distortions and inaccuracies in witnesses’ memory that result from the memory process itself.
18. This Report is comprised of six sections and two Appendices which are integral parts of the Report.
19. Section II ‘What existing scientific research tells about memory and eyewitness evidence’ refers to the work of Professor Maryanne Garry and summarises the report co-authored by Dr Ula Cartwright-Finch and Dr Kimberley Wade (at Appendix 1) that sets forth the contemporary science with respect to memory, with a particular focus on aspects pertinent to witness evidence.
20. Because much of the existing scientific research with respect to witness memory centres on criminal cases in which witnesses typically are asked to recall what they saw and heard with respect to a particular incident, the Task Force pondered whether the findings of those studies – that human memory is subject to distortion by post-event information and other factors – also applied to disputes in international arbitration which normally do not centre on one critical event (i.e. the ‘commission of the crime’) but rather involve a sustained interaction between parties (witnesses) over time, including communications in person and via other means. Another feature of international arbitration is that witnesses are only asked to give evidence by a party if their evidence supports the case being advanced by that party. Witnesses who are entirely neutral to the parties are rare.
21. To address these concerns, the Task Force benefitted from the expertise of Dr Kimberley Wade at the University of Warwick, who undertook a series of field studies designed to determine the vulnerability of memory in the context of a commercial dispute not atypical of the disputes that are routinely resolved in arbitration.
22. Section III ‘Do the same witness memory issues arise in international arbitration?’ summarises the key findings of Dr Wade’s research (at Appendix 2) that reveals that the memory of witnesses in international arbitration is subject to the same distorting effects research has proven exists in other contexts.
23. Section IV ‘Accuracy of witness memory and its significance in international arbitration’ examines the purposes for which witness evidence is used in international arbitration and evaluates the contexts in which accuracy of witness memory is and is not important. It draws in great measure upon the views of the Task Force members based on their experiences.
24. Witness memory can be affected by a variety of factors, including post-event information and bias (such as pro-employer). Multiple actors in the course of a dispute, including in-house counsel, outside counsel, peers and supervisors, can impact a witness’ memory process, so each of the participants in the arbitral process can help to reduce the distortion effects.
25. Section V ‘Measures that can be taken to improve the accuracy of witness memory’ sets forth specific suggestions about how in-house counsel and outside counsel can carry out the fact-gathering and fact-presenting process so as to reduce the risk of memory distortion. It also describes certain steps that an arbitral tribunal may consider taking in appropriate cases, both to reduce the distortions of witness memory and in evaluating witness evidence in the light of any potential distortions.³
26. Section VI sets out the Task Force’s ‘Conclusions and Recommendations’. As is stressed in several places in this Report, the measures described in Section V are not intended to be one-size-fits-all. In essence, the suitability and utility of each suggestion can be evaluated only in the context of the specific circumstances. The efforts to keep witness memory pristine must be tempered with practical considerations. Sometimes, for example, it is not practical to interview witnesses individually. As well, sometimes post-event information, such as reviewing a document, can refresh or reactivate a witness’ memory accurately. In addition, different arbitrators will have different philosophies about the role and importance of witness evidence and the extent to which they should become involved, if at all, in discussions with the parties or the witnesses about the way in which witness evidence is prepared. Accordingly, the Task Force recognises that some arbitrators will not consider it appropriate to take some of the steps suggested – but they are included in the Report so that they can be considered by all and used selectively by those who consider they may be helpful.
27. The Task Force is of the view that when dealing with something as complex and intangible as how to obtain the maximum value from witness evidence, there is no single 'best' practice. Given the variety of issues relative to witness memory (and therefore witness testimony), tribunals and parties need to evaluate what measures are called for in the particular circumstances, which measures are worth the effort, and how important the risk of memory distortion is to the outcome in light of the facts of the case.
28. Thoughtful analysis is required to avoid under - or over- correcting. For example, some practices, such as having counsel prepare a witness statement, can have a distorting effect upon the witness’ memory but, nevertheless, have a significant benefit in helping a lay witness communicate to the tribunal the witness’ understanding of the facts. All measures offered for thought in this Report are also subject to overriding legal requirements of the particular jurisdiction involved.
29. Once the process of determining what happened in a dispute is based on a reconstruction of events after the fact, it is destined to be imperfect. The fact that witnesses’ memories are imperfect does not mean that their testimony is not valuable or at times vital. The aspiration should be to reduce the imperfections to the extent reasonably possible in order that the decision rendered can be just, based on a reasonably close approximation of what in fact happened. The measures recommended in this Report are one step in that direction.
30. The value of witness evidence is affected by many factors, memory being just one. Factors such as cultural perceptions, manner of presentation, language, and cognitive biases by arbitrators, counsel and witnesses, all bear on the value of witness evidence. In the course of the Task Force’s work, Task Force members made salient observations with respect to these other factors.
31. While the Task Force was tempted to study such factors as well, after careful consideration, it concluded that those subjects were beyond the Task Force’s remit which was specifically to study and make recommendations about the effect of memory on the value of witness evidence. As a result, the many interesting and valuable observations made by Task Force members with respect to such other factors are not included in this final Report, but that is not intended to imply that witness evidence is not affected by factors other than memory.
32. The Task Force in fact considers that the value of witness evidence could be enhanced through further study of such other factors, quite apart from memory. The Task Force would encourage ICC and the arbitral community in general to consider studying such other ways to enhance the value of witness evidence in the arbitral process.
33. While the many salient points raised by Task Force members with respect to other factors that affect or might affect the value of witness evidence are not included in this final Report, one – cross-examination – warrants mention. Task Force members commented on the effect of cross-examination on witness evidence, there being a range of views on the subject. As the Task Force’s remit is the effect of memory on the value of witness evidence, the Task Force sought to evaluate the effect of cross-examination on witness memory – not the effect of cross-examination on the value of witness evidence in general.
34. Dr Wade for sharing her knowledge explained:

‘[M]any studies have focused on the accuracy of eyewitness accounts and the factors that increase or decrease accuracy, [but] relatively fewer studies have examined memory accuracy as people recount their experiences on different occasions. Given there are few studies into the effect of cross-examination style questions on memory accuracy, it would be premature to make any practical recommendations.’⁴

As a result, the Task Force makes no recommendations with respect to cross-examination.

II. What existing research tells us about memory and eyewitness evidence

1. The starting point for this Task Force was to understand the existing research that has been undertaken into the fallibility of human memory insofar as it is relevant to witnesses. That research has largely focused on the memory of eyewitnesses, and the potential impact of erroneous memories on criminal justice.
2. Once that research was understood, the Task Force could move on to consider its potential impact on witness evidence in international arbitration. That consideration would necessarily include an assessment of whether new research would assist in confirming whether the conclusions drawn in relation to eyewitness memory are equally applicable to the memory of witnesses who give evidence in commercial disputes – typically business people who are recalling meetings, discussions and other events relating to a business transaction.
3. At the outset of its work, the Task Force was introduced to Professor Maryanne Garry, at the School of Psychology of the University of Waikato in New Zealand. Professor Maryanne Garry and her co-author Eryn Newman describe scientific studies that lead to the conclusion that ‘memories are not an objective, unyielding imprint of the past, but a subjective, pliable patchwork of experience, thoughts and daydreams’.⁵
4. Professor Garry and Eryn Newman’s chapter describes studies which investigate the impact that exposure to post event information (‘PEI’) can have on a person’s memory. In particular, it looks at the impact that misleading PEI can have on the way in which a witness recalls a specific event. The following short excerpts from the chapter provide a snapshot of some of the key findings:

‘Over the last four decades, hundreds of experiments conducted in laboratories all over the world have demonstrated that people unwittingly adopt misleading suggestions as their own memories. Although here we focus on how information after an event can shape people’s memory reports, information that people encounter before an event can also shape memory.’

‘Talking to co-witnesses after an event, viewing media reports, or retelling an experience to friends or officials are all opportunities for PEI [post event information] to creep into memory reports.’

‘This line of research suggests that simply sharing our experience with someone else can be an opportunity for misinformation to invade memory.’

‘[T]his research tells us that misinformation doesn’t just alter details in our memory, it can add information to memory that was never there in the first place.’

‘[G]uessing and describing can act as misinformation, so that details people first offer as guesses can later corrupt memory for the event’

‘[A] memory for the event becomes weaker, it becomes more susceptible to misleading PEI.’

‘When the misinformation messenger seems particularly credible, inaccurate PEI is more likely to creep into people’s memory. Findings from co-witness research fits with this idea: witnesses who believe that another witness saw an event for longer are more likely to adopt that witness’ inaccurate suggestions.’

‘[I]f people can track the sources of their memories easily and they are told to watch out for errors, they can better avoid misleading suggestions. In general, warnings are most effective when they occur before the PEI, because people can carefully scrutinize subsequent information and protect themselves from misleading suggestions.’

5. The scientific research that these short excerpts relate to suggests not only that human memory is very unreliable, but also that a witness’ exposure to pre and post event information regarding an event can add to, detract from or change a memory of the event.
6. The potential relevance for the witness evidence that is presented in international arbitration is clear. The way in which witness evidence is currently prepared and presented in arbitration proceedings includes multiple opportunities for a witness to be exposed to PEI, including potentially misleading PEI. Does that lead to a risk that the evidence of honest witnesses may be less reliable than arbitral tribunals currently assume?
7. As a first step in answering that question, Dr Ula Cartwright-Finch⁶ and Dr Kimberley Wade prepared a report on three specific areas of research that are particularly pertinent to witness evidence in international arbitration. This report entitled ‘Selected Review of Scientific Literature relating to Memory and Witness Evidence’ (Appendix 1 to the Task Force’s Report) focuses on three areas:

• How the specific wording of a question can change the way a witness replies.
• The way in which a witness may be influenced by information received after an event (the so-called ‘misinformation effect’), including (i) situations where distortions are introduced into the witness’ own memory reports; (ii) misinformation arising in the context of discussions between co-witnesses; and (iii) the creation of entire false memories.
• How the act of retelling a story from a particular perspective can change a witness’ memory.

8. Readers are encouraged to read the report of Dr Cartwright-Finch and Dr Wade in full. This section follows with a brief summary of some of the studies described in the Report and the conclusions that can be drawn from them.

A. The impact of phrasing on responses to questions

9. The preparation of witness evidence in arbitration typically involves a witness being questioned or interviewed by a lawyer who represents the party that wishes to rely on that witness evidence. That lawyer may go on to prepare a first draft of a witness statement based upon the answers given by the witness. Studies which have looked at the impact of phrasing in questions on the memory recall of a witness are therefore of interest.
10. Studies have shown that qualifying descriptors in a question can heavily influence the answer given. For example, in one study, participants who were asked to estimate ‘How long was the movie?’ answered with an average of 130 minutes vs those who were asked ‘How short was the movie?’ who responded with 100 minutes on average.
11. Similarly, in a separate study, the question ‘Do you get headaches frequently, and if so how often?’ tended to produce a different answer to the question ‘Do you get headaches occasionally, and if so, how often?’ Those answering the first question reported on average 2.2 headaches per week, whereas those answering the second question reported only 0.7 headaches per week.
12. The conclusions drawn from these and similar studies is that the specific wording used in witness interviews is important. Changing just one word within a question can materially change the evidence that a witness recounts.

B. The misinformation effect

13. The misinformation effect describes a phenomenon where typically misleading information which participants are exposed to after an event interferes with or impairs their original memory of that event.
14. Potential witnesses in arbitration proceedings may well come into contact with such misinformation. This might come from management, colleagues or fellow witnesses, or it might come from in-house or external counsel who discuss the witness evidence in preparation of the case. The potential effect of misinformation on witness memory is therefore of interest.
15. One way in which misinformation can be transmitted is from another witness. For example, studies have shown that where two subjects viewed somewhat different versions of an incident (e.g. highly similar scenes of a mock crime that are filmed from different camera angles) and discussed the events subsequently, they unwittingly transmitted misleading information to each other, which can substantially alter a witness’ memory and report. Even more, the report by Dr Cartwright-Finch and Dr Wade informs that people tend to incorporate details of other witnesses’ memory reports into their own memory reports, even if that information contradicts what they have observed.
16. The relevance of this to international arbitration is evident. Disputes in arbitration normally involve multiple persons with knowledge of the pertinent facts (witnesses), and as part of the normal process of dealing with problems, those persons communicate and interact with each other, both before and after counsel are on the scene. Beyond that, it is not unusual for witnesses to be interviewed or debriefed together.
17. Of particular relevance to international arbitration is a recent large cross-cultural study, reported by Dr Cartwright-Finch and Dr Wade, that demonstrates the effect of co-witness discussion on memory. This study included pairs of participants from ten countries (Brazil, Canada, Colombia, India, Japan, Malaysia, Poland, Portugal, Turkey and the United Kingdom) who each watched slightly different versions of the same film and then discussed the conflicting details they had viewed. Participants often reported details on the memory test that they had not witnessed. The key takeaway for international arbitration was that culture did not keep witnesses from incorporating details of other witnesses’ memory reports into their own memory reports, even if that information contradicted what they had observed. Regardless of whether that incorporation was the result of a memory distortion or of witnesses’ trusting reliance in their co-witness, the study is relevant to international arbitration.
18. The misinformation effect has been replicated across thousands of studies in a wide variety of contexts and using a variety of different stimuli.
19. A witness’ exposure to misinformation may come about in different ways, one of which is in the form of questions that are put to the witness. For example, participants in a study who were asked to recall ‘Did you see the broken headlight?’ were far more likely to recall a broken headlight than participants who were asked ‘Did you see a broken headlight?’. The simple use of the definite article which implied that there was a broken headlight in the video clip that the participants had viewed was sufficient to affect the participants’ recollection.
20. Studies show that the incidental presentation of misinformation can strengthen the memory changing effect: participants in a study who had been shown a video clip where no school bus had been present were more likely to respond positively to the question ‘Did you see the children getting on the school bus?’ than they were to the question ‘Did you see the school bus?’.
21. Studies have also shown that misinformation can have the effect of overwriting an existing factual memory in addition to supplementing memory with additional (inaccurate) details. For example, participants who had seen a video of a car accident next to a ‘Stop’ sign were asked questions which included reference to the fact that the car had been stopped next to a ‘Yield’ sign. When later asked to identify the correct image from the video, 59% chose the image with a ‘Yield’ sign and only 25% chose the correct image with the ‘Stop’ sign that had featured in the video.

C. False memories

22. The studies referred to above examined memory in the context of a participant passively witnessing an event such as a road accident. A series of striking studies extended these results to show that it is possible, in some circumstances, to lead participants to ‘remember’ entire fabricated events which have happened to them personally, so-called ‘false memories’.
23. For example, participants who were reminded of true events from their childhood were, over time, led to also recall a fabricated event (getting lost in a shopping mall) by the inclusion of questions about that fabricated event in successive interviews with the researchers. Over the course of multiple interviews, some participants reported the episode with greater detail, confidence and embellishment.
24. Misinformation can therefore be used to implant entire events within a person’s memory. A further example of this was a study that involved the digital manipulation of old photographs. By showing adult participants a photo which had been digitally enhanced to show them riding in a hot-air balloon, some participants were led to report complete or partial memories relating to a totally fabricated hot-air balloon ride. This further extension of the false memory effect has significant implications given the widespread availability of digital tools designed to doctor photographs and documents.
25. False memories can arise from manufactured evidence. Dr Cartwright-Finch and Dr Wade report on one experiment in which a subject participating in a gamin exercise next to another subject was asked to corroborate that he had seen the other subject cheating. Those subjects shown a false video were more likely to provide corroboration (of something they had not seen and had not happened – cheating) than those who had not seen the video. The saying that ‘seeing is believing’ is especially fraught with risk in the modern digital world.

D. The impact of retelling on subsequent recall

26. Research has been conducted to examine how taking a particular perspective after an event is encoded into memory can bias later recall of the event towards a biased perspective. Of course, in the context of a commercial dispute, a witness is very often engaged in telling his or her story from a particular perspective – most commonly that of a claimant or respondent by whom they are often employed. The potential impact of this biased retelling on the witness’ memory is therefore relevant to the typical arbitration scenario.
27. Participants in one study were given a story about their first week of a new year at college and the interactions they had with two roommates. One group were asked to write a letter complaining about one of the roommates. Another group were asked to write a letter recommending one of the roommates. A third (neutral) group were just asked to write about the roommates.
28. When their recall was later tested, the participants’ recollection of the original roommate story was materially influenced by the biased retelling that they had been asked to undertake when writing the letters. Participants remembered more about the roommate they had written about, and also made more biased errors about that roommate. The neutral group did not demonstrate the same errors in the recall task. Thus, participants in this study used the original story to suit their task, and that act of perspective-taking affected their subsequent recall of the original story.

E. Maximising the completeness and accuracy of witness memory reports

29. Dr Cartwright-Finch and Dr Wade report on studies that show how steps can be taken to ameliorate distortions on memory. For example, studies indicate that people are much less likely to experience memory errors if they can retrieve information in memory that helps them to identify any details that are distorted or have been suggested to them.
30. Dr Cartwright-Finch and Dr Wade note a number of steps that can be taken to ameliorate distorting effects, such as:
1. Have witnesses provide a complete account of the witnessed event immediately after it has happened. Doing this will reinforce the memory for later recall and reduces the chances of subsequent contamination.
2. As discussed above, the phrasing of a question can impact the witness’ answer and affect memory subsequently. The authors cite the use of open-ended questions rather than specific or potentially leading questions as one technique to reduce distorting effects.
3. Where a witness is unsure or tentative about an answer, the questioner can have a distorting effect upon subsequent memory; interviewers should not reinforce tentative or unsure responses as this can inflate a witness’ confidence in testimony they are actually uncertain about.
4. Studies show that warning witnesses that they have been misled – after being misled but prior to having their memories tested – can reduce memory errors; reminding witnesses that their task is to report only their own personal knowledge can reduce memory distortion.
5. Studies also show that witnesses have more accurate and less distorted memory recounts when encouraged to identify the source of their knowledge.

F. Conclusions

31. The Task Force was persuaded that the existing research raised two important questions in the context of international arbitration:
• Is there a risk that in some cases, witness evidence which is based upon witness memory is not as reliable as a tribunal might have assumed?
• If so, are there steps that can be taken to improve the reliability of witness evidence?
32. In order to answer the first question, the Task Force felt that further scientific research would be helpful. Specifically, the Task Force wanted to test whether similar results would be produced by a study which tested witness memory in the context of a commercial dispute which might typically be the subject of international arbitration proceedings. The next section of this Report describes the study that was undertaken and its findings.

III. Do the same witness memory issues arise in international arbitration?

1. In order to test memory in a commercial setting, the Task Force and Dr Wade developed a witness memory experiment. The experiment was designed as an online study. A total of 316 adults, working across a broad range of industries and roles, participated in a standard witness memory experiment. The experiment confirmed that memory mistakes also occur in a commercial setting typical of arbitration proceedings. Details of this experiment are set out in Dr Wade’s experimental report (Appendix 2).
2. For purposes of the experiment, participants initially read about a contractual agreement between two companies that ultimately led to a dispute. Participants first read a summary of the contractual relationship between a printing company and an industrial flooring company. The printing company contracted the flooring company to replace the floor in their printing plant. Next, participants studied the contract and purchase order created for the works before reading a transcript of an in-person meeting that took place between the two companies not long before the works commenced.⁷
3. In order to examine how participants’ (biased) perspective might affect their subsequent memory of key events, some participants were instructed to imagine they were the Managing Director of the printing company and others were instructed to imagine they were the Managing Director of the flooring company. Some (control) participants were not told to imagine anything. Further misleading information was introduced in the course of the experiment.
4. After a short delay, participants were informed that a dispute has arisen between the two companies around two issues.
5. First, the flooring company’s invoice included a surcharge that the printing company claimed they had never agreed to pay. In fact, no such agreement for the payment of a surcharge was reached at the meeting.
6. Second, the new tile floor started to crack not long after being installed; the printing company believed that the tiles were not fit for the purpose required by the contract. The contract said nothing about the printing machines being rolled around the floor which would add to the load; however, the printing company argued that a representative from the flooring company was indeed aware that the machines moved around because he saw the machines moving during a meeting at the printing plant. The flooring company claimed that they were not aware that the printing machines would be rolled around; the facts of the case made it clear that the participants on the part of the flooring company could not observe machinery being moved around while they visited the printing company.
7. On this second issue, some of the participants, in addition to imagining being managing directors of either company, received additional (misleading) information in form of a memo by their own company’s in-house counsel. While the in-house memo for the flooring company group suggested that the witnesses could not observe machinery being moved around, the memo provided to the printing company group suggested the opposite, i.e. that the staff of the flooring company could observe machinery being moved around (as stated, no such observation was possible).
8. Participants were then informed that the two companies agreed to arbitration and that they (each participant) had been called upon as a fact witness.
9. Participants then answered a series of questions about their memory of events. Most importantly, they were asked two questions about the key issues set out above.
10. First, participants were asked whether the printing company agreed to a surcharge at the in-person meeting. While the answers appeared to fit the prediction that participants would be biased towards their own company, the analyses showed that there was no (statistically) significant association between the company that participants imagined they worked for and their response to this question. The available data, however, suggests that this is because participants found the issue too complicated.
11. Second, participants were asked whether a representative from the flooring company viewed the machinery on the factory floor being moved around. In regard to the second issue, the results revealed that instructing participants to imagine that they worked for one of the two companies and exposing participants to (biased) post-event information in the form of an in-house memo, influenced how they responded on the witness memory test.
12. Among the participants who imagined being the Managing Director of the flooring company, 66% responded that machinery could not be observed being moved around. This number increased to 78% for those participants who were further exposed to the in-house memo (suggesting that machinery could not be observed being moved around). For the participants who were asked to put themselves in the shoes of the Managing Director of the printing company, only 43% stated that machinery could not be observed being moved around; this number dropped to 29% for those participants who were further exposed to a misleading in-house memo (wrongly suggesting that machinery could be observed being moved around. It is interesting to compare these percentages with the control group containing participants who were not affiliated to either company and did not receive a biasing memo. In the control group, 50% responded that the machinery could not be moved around. This compares to 78% for the flooring company group that received a biasing memo in favour of the flooring company, and to 29% for the printing company group that also received a biasing memo in favour of the printing company.
13. The results from Dr Wade’s study were consistent with those found in witness memory studies conducted in the criminal context. In particular, biasing people in favour of a particular company and exposing them to suggestive post-event information affected their memory reports. Each of these factors led participants to recall details in a way that better supported their own company’s case. Taken together, these findings illustrate that just like witness memory in the criminal context, witness memory in business settings is liable to error.

IV. Accuracy of witness memory and its significance in international arbitration

1. The existing research (summarised in Section II of this Report) and Dr Wade’s study (summarised in Section III) show that a high level of interaction between a witness and other sources of information can lead to distortions of memory.
2. In modern arbitration proceedings, such interaction starts when potential witnesses first receive information about what ultimately turns out to be a material fact. For instance, memory may be influenced by discussions between the potential witnesses as co-workers. Such interaction continues when the potential witness provides information to in-house counsel or external counsel for the purposes of drafting submissions. The level of interaction even increases when it comes to the stage of preparing witness statements, preparing - to the extent permissible - witnesses for the evidentiary hearing, and finally the examination of witnesses. If one applies the existing research to international arbitration proceedings, there undoubtedly is a high risk of memory distortions.
3. On this basis, it is tempting to recommend reducing the level of interaction, in order to reduce the level of memory distortions on the part of fact witnesses. However, such a recommendation would be premature for two reasons.
4. First, fact witness evidence in international arbitration is not just about memory. Parties also adduce fact witness evidence for other purposes in which memory plays only a limited role, or no role at all (see paragraphs 4.6 to 5.4 below).
5. Second, some of the procedures that entail the risk of memory distortions are beneficial for other reasons. In the view of the Task Force, it is therefore necessary to weigh the advantages and disadvantages of such procedures in order to avoid recommending a cure that is worse than the disease. This balancing exercise is discussed in Section VI ‘Conclusions and Recommendations’ at the end of this Report.

A. Fact witness testimony is not only about memory

6. The research referred to in Section II was predominantly carried out in the area of criminal law, i.e. with a focus on eyewitnesses. In a criminal context, the essence of fact witness testimony (a witness to a crime) is about what witnesses have seen or have heard. Naturally, avoiding distortions of memory is highly important in such a setting. In international arbitration, the emphasis is usually different. Proving disputed facts that are material to the outcome of the case is not the only purpose of fact witness evidence.
7. As a discussion at an early meeting of the Task Force demonstrated, a party to arbitral proceedings and its counsel might decide to present the evidence of a particular witness to an arbitral tribunal for a variety of reasons. To give one commonly quoted example, witnesses can be used to give general background information and to ‘set the scene’ in which the dispute arose. When used in that way, it is quite possible that the evidence of a witness will not concern disputed facts at all.
8. For example, a joint venture dispute where the issue is whether one joint venture partner had performed its obligations to support the joint venture in accordance with the terms of a joint venture agreement. The CEO of the claimant describes in his/her witness statement the nature of that claimant’s business, how the business had grown over the years and what its hopes and aspirations were when entering into the joint venture.
9. These are not matters which turn on the CEO’s memory of specific events, and it is quite possible that they are not disputed by the respondent. The CEO’s evidence is not material to the arbitral tribunal’s determination of the issues in the case concerning an alleged failure of the respondent to perform specific obligations under the joint venture agreement. Nonetheless, the claimant and its counsel may well feel that it is helpful to include that material in the CEO’s witness statement in order to give the arbitral tribunal what it feels is relevant background to the case and to its position. They may also feel that it is helpful to their case for the tribunal to be able to meet, hear from and ask questions to a senior member of the claimant’s management.
10. Recognising that in international arbitration witness evidence is used for purposes beyond proving disputed facts, the Task Force decided to explore in more detail what those uses were.
11. To this end, the Task Force circulated a questionnaire to members of the ICC Commission on Arbitration and ADR. These members include arbitrators, counsel, and user representatives from around the world. For the purposes of this Report, the questions in the questionnaire included:

• What are the purposes for which fact witness evidence is actually used in international arbitration?
• Are there purposes for which fact witness evidence is used which you do not consider useful?
• Could any of these purposes be served by other means?

12. These questions are addressed in Section IV(B) below, on the basis of the responses received from Commission members (a selection of which are quoted).

B. Different purposes of fact witness testimony

1) Proving disputed facts

13. The principal function of witness evidence was summarised succinctly in the following response:

‘[T]o prove disputed facts, in particular those that cannot be proven by written evidence.’

14. This is uncontroversial. Where a party needs to prove a fact in order to make good its claim or defence, and that fact cannot be proven by documentary evidence, then witness evidence (if available) may be essential. There is unlikely to be any alternative. Of course, if the fact that needs to be proved arises out of historic events, then the accuracy of the memory of the witness of those events is important.

2) Explaining documents

15. Parties sometimes choose to present witness evidence on matters which are also addressed in documentary evidence. The reasons for this can vary. The following examples were given:

‘To provide a context for the contemporaneous documents.’

16. It may be helpful for an arbitral tribunal to understand the context in which important documents have been created or sent. For example, when reading a sequence of emails which set out a discussion between the parties on a particular topic, it may be relevant to hear evidence about a telephone conversation that took place on the same topic between the same parties in the course of that exchange. Whilst what was said on the telephone call may not be disputed, if the content of the conversation helps to explain the position a party has taken in a later email, that party may well feel that it is necessary to produce witness evidence of the conversation.

‘Sometimes the plain language of a document can be misleading or can have several meanings. A witness can explain what he/she meant when he/she wrote (or received) a document.’

17. Whether witness evidence about what a witness meant when he or she wrote or agreed to wording in a document is relevant is likely to depend on a couple of factors.
18. Having a witness explain what he or she meant when using particular words in an email or letter may well help the tribunal understand what the witness was trying to say in that communication.
19. It is worth noting in this context that the relevance or admissibility of a witness evidence about his or her subjective intention when agreeing to certain ambiguous language in a contract may well depend on the governing law. National laws differ in the approach that they take to establish the true meaning of an ambiguous provision in a contract.
20. Whilst witness evidence to explain documents may be considered useful in some cases, it is worth noting that amongst the ways in which witness evidence is used which responders to the questionnaire felt were not useful, high on the list was repetition of documentary evidence. For example:

If relevant and material facts are already well documented by written exhibits, a repetition of the respective content by witnesses may not be very useful, but only time consuming and costly.

3) Providing context and ‘telling the story’

21. Many members of the Task Force and Commission members who responded to the questionnaire stated that they use witness evidence to tell their or their client’s story.

‘To tell the party’s story on a more personal / engaging level.’

‘Filling in background and painting the tapestry of the factual matrix – there is often a lot of context behind a dispute that is not reflected in the documents.’

‘Increase credibility of the party’s position.’

22. Some responders said that it was important to be able to put a human face on information contained elsewhere in the documents and that witnesses could be used to elicit sympathy from the tribunal, for example by creating an impression that the party is a reputable company that always honours its commitments.
23. It is certainly common in international arbitration (perhaps increasingly so) for witness statements to include sections (sometimes lengthy) which provide background information from the perspective of that particular witness. Those responding to the Task Force’s questions as counsel seemed to consider this to be more important and more useful than those responding as arbitrators. One arbitrator, however, commented:

‘Witness evidence could be used to get an impression of the persons involved in the disputed fact patterns – all with the purpose to have a sound basis coming as close as possible to the "truth" to render an award.’

24. Whilst using witness evidence to tell a party’s story may be a popular technique, many counsel and arbitrators who responded to the questionnaire expressed frustration at the use of witness evidence to argue a party’s case. A commonly held view is that it is inappropriate for witness statements to be used to advocate a case, make legal arguments, offer opinions on legal interpretation and generally to address matters which are properly the domain of legal submissions. One arbitrator put it in these terms:

‘Sometimes testimony by a party’s officers is submitted as evidence with little purpose beyond explaining that party’s position with respect to the dispute. That purpose is, in my view, better left to the party representatives and their written submissions or oral presentations. Similarly, the meaning and significance of documentary evidence submitted by the parties should primarily be explained to the tribunal in written or oral submissions. It is not always necessary to have witnesses explain the meaning of a document, especially where its text is clear on its face.’

25. Frustration was also expressed at witness statements that set out irrelevant information about the witness’ personal background. A number of arbitrators also commented that witness evidence is used by some parties simply because the opposing party has done so. It was described as ‘a ritual rather than a means to put a spotlight on really decisive factual issues’. One arbitrator commented that witness evidence is used ‘all too often’ to obstruct or delay the proceedings by submitting witness testimony on irrelevant or immaterial facts. Again, however, the question arises as to whether such perceived lack of focus could be avoided if a tribunal shares its preliminary views on what is relevant and material with the parties at an early stage. Another arbitrator was critical of witness statements that were ‘used by lawyers to provide a slanted narrative description’.

4) Providing technical explanations

26. In technical cases, such as engineering or infrastructure disputes, witnesses often give testimony on the technical aspects of a dispute. While such explanations could in theory be left to expert witnesses, the parties’ employees involved in such disputes often have a knowledge that equals the knowledge of expert witnesses, or might even be superior in case of highly specialised industries. Such witnesses, for instance, can provide useful explanations to a tribunal about a project in general, the functioning of a piece of machinery or even the root cause of a defect. In such cases, fact witness testimony is more about the witness’ knowledge and experience; such that memory distortions are less relevant.

C. Conclusions

27. The responses to the questionnaire showed that there was no common ground as to whether the various purposes for which witness testimony is actually used in arbitration are all valid. For instance, some responders believed that providing context, also in the form of information that is not strictly material and relevant, is a valid purpose; at the same time, other responses indicated that such circumstantial information would render witness testimony less efficient.
28. To a certain extent, but not solely, these differences corresponded with legal backgrounds, i.e. whether responders had a common law or a civil law background. Generally, responders with a civil law background believed that fact witness testimony would benefit from being limited to facts that are disputed and relevant to the outcome of the arbitration. Responders with a common law background tended to accept a wider approach to fact witness testimony as useful. Another soft dividing line could be discerned between responders who answered from an arbitrator perspective and those who answered from a counsel perspective. Maybe not surprisingly, arbitrator answers favoured a more focused and narrower approach to witness testimony than counsel did. This held particularly true for witness statements.
29. Whilst the responses to the questionnaire provided a fascinating insight into the different reasons that parties and arbitrators consider witness evidence to be useful, it is not this Task Force’s mandate to evaluate the usefulness of the mentioned purposes.
30. Having regard to this Task Force’s topic, i.e. memory distortions, the following conclusions can be made.
1. The degree to which the probative value of witness evidence depends on the accuracy of memory varies in light of the different purposes that parties pursue by adducing fact witness evidence. If, for instance, a witness provides certain background information to a project, typical business procedures in his or her company, the mechanism of a piece of machinery, etc., such information is rather based on experience, education or knowledge. If, for instance, a witness describes the process of effecting payments within their company, e.g. how invoices are scrutinised and who has authority to greenlight payment, such testimony is about knowledge rather than memory. It is not about what a witness has perceived, i.e. has heard or has seen, at a specific point in time. For such witnesses, it is unnecessary to consider steps and procedures that reduce the likelihood of memory distortions.
2. On the other hand, if fact witness testimony is about specific events, such as statements that were made during a meeting (in particular a meeting for which no written record exists) the accuracy of memory is likely to be important.
31. In light of these conclusions, the Task Force therefore considered that it would be helpful to set out steps that could be taken to ensure that witness memory is preserved as best as it can be. This is the focus of the next section of this Report. Such steps can be considered in cases where the accuracy of witness memory is likely to be material to the outcome of the case.
32. However, the Task Force also concluded that such steps may not be relevant, feasible or practical in many cases. Before embarking on steps which seek to preserve the accuracy of witness memory, it is first necessary to decide whether the accuracy of witness memory is likely to be a material factor in deciding the case. If witness evidence is being produced in the case for some of the reasons described in this section which do not depend upon witness memory, then the steps discussed in the next section of this Report are unlikely to be warranted.

V. Measures that can be taken to improve the accuracy of witness memory

1. This Section of the Report describes certain measures that can be adopted to preserve or increase the accuracy of witness memory.⁸
2. There was a wide consensus among the Task Force that the options listed in this Section must not be understood as ‘rules’ or ‘best practices’ that the Task Force would recommend be complied with. Rather, it is an open list that practitioners can consider and select from as appropriate.
3. The measures described are not intended to be one-size-fits-all. Given the variety of issues relative to witness memory (and therefore witness testimony), tribunals and parties need to evaluate what measures are called for in the particular circumstances, which measures are worth the effort, and how important the risk of memory distortion is to the outcome in light of the facts of the case. Thoughtful analysis is required to avoid under- or over-correcting.
4. Section V categorises measures into two groups. The first group is comprised of measures that can be taken to reduce distorting influences and their effect (A). As it might not be possible to eliminate distorting effects completely given the very nature of human memory, the second group is comprised of steps that can be taken to allow the tribunal and the parties to identify and weigh the distorting influences that might exist, and thereby take them into account when weighing the witness testimony (B). These steps are set out by reference as to whom they are aimed at – in-house counsel, external counsel, or arbitrators.

A. Measures that can be taken to reduce distorting influences and their effect on witness evidence

1) In-house counsel

5. In-house counsel are often the ‘first responders’ on the scene when a dispute arises, and therefore have the opportunity to take steps to mitigate or eliminate the factors that can distort the memory of witnesses. In-house counsel could:
1. Establish procedures for teams to keep contemporaneous written or oral notes of issues being discussed at the time the relevant events unfold, especially in a potentially or actually contentious situation.
2. Emphasise to witnesses the importance of their own recollection being presented to external counsel. It is likely to be in the party’s interests for counsel to be able to assess the witnesses’ evidence fully, including the points which are unhelpful to its case.
3. Where practicable, meet with likely witnesses individually, and avoid meeting with likely witnesses in groups to discuss the case, either internally or with external counsel, in order to avoid memory contamination.
4. Strive to have present in witness interviews only those persons needed to conduct the interview effectively.
5. Avoid setting out a 'party line' to prospective witnesses which may have the effect of modifying their recollection of events so as to match the 'party line'.
6. Discourage witnesses from needlessly discussing the matter amongst themselves. If witnesses do wish to do so, in-house counsel may be present to help ensure that the witnesses’ memory is not impacted by such discussions.
7. Involve external counsel early in the process (once a matter becomes contentious), then identify and preserve witness evidence while it is still fresh and authentic. This may include facilitating early witness interviews with potential witnesses and/or providing to all prospective witnesses guidelines aimed at reducing the risk of confabulation/false memory creation.

2) Outside counsel of party presenting witnesses

6. Outside counsel often play a central role in the gathering of evidence once a dispute has arisen and, in the process of gathering facts, interact with witnesses in ways that can affect the witness’ later memories. Outside counsel may consider the following measures:

(i) Conducting interviews

7. Timing. Interview witnesses at the earliest opportunity. The memory is likely to be more accurate (i) the closer to the event that is being recalled and (ii) the less it has been exposed to contamination from other sources of information.⁹
8. Setting of interviews. Where feasible, interview witnesses individually – not in a group –to minimise memory contamination, i.e. to avoid having the memory of a witness change based on discussions with other individuals. In cases where outside counsel becomes involved only at a later stage, after the start of the dispute, establish how much prior discussion and contact there has been among the witnesses in order to gauge the extent the recollection may have already been modified.¹⁰
9. Keep an accurate record of interviews. Consider having two persons conduct the interview, with one person being the primary interviewer and the other keeping detailed notes. Also consider keeping an audio or video recording (but to be clear, the Task Force is not proposing that witness interviews be recorded or taped as a general practice).¹¹
10. First interview of the witness: Conduct the first interview with the following guidelines in mind (these guidelines are for consideration in each case and may not be appropriate in every situation):
1. Put the witness at ease and in a condition to communicate the most accurate version of the facts by explaining some basic guiding principles:
• Remind the witness that it is normal to have forgotten details and events. If the witness does not remember, they should simply say so. If the witness is making assumptions (e.g. on the basis on minutes of meetings or a calendar entry, the witness assumes that they attended a particular meeting), the witness should state that it is an assumption that is being made.
• Ask witnesses to distinguish between what they remember rather than what another person may have told them or what they may have read. If witnesses relay information outside their personal scope of memory, they should state the source of such information.¹²
2. Assuming it is true, take steps to assure that telling the truth will not result in personal consequences to the witness.
3. Inquire about the discussions and correspondence that the interviewees already had regarding the memory they are being asked to recall.
4. Reconstruct the trajectory of the witness’ experiences since the event being recalled occurred. Who did the witness communicate with, how many times, under what circumstances, what documents did the witness see or generate, etc.¹³
5. Strive to make the questions unbiased and open-ended and review them in advance more than once to assure their 'neutrality'.¹⁴ Unless carefully drafted, questions can be leading or impliedly suggest an answer unintentionally.
6. Use neutral language during interview and avoid qualifying descriptors as these may influence the witness’ recalling of the memory. For example, do not ask ‘How aggressively did the Respondent’s manager react?’ or ‘How stubbornly did they resist your request?’ as the qualifying descriptors ‘aggressively’ or ‘stubbornly’ may impact the witness’ response. Instead, use more neutral language: ‘How did the discussion at the meeting progress?’.¹⁵
7. Avoid intervening and influencing a witness’ answer. Specifically avoid interruptions, inappropriate questioning styles, leading questions/responses and interrogative pressure. Let the witness tell their story first before conducting more in-depth questioning.¹⁶
8. Avoid giving feedback to the witness on their answers, irrespective of whether such feedback may be negative or positive (e.g. avoid stating that you agree or disagree with the witness’ answers, or that the evidence/response given to the question conflicts with the pleaded case or is otherwise unhelpful to the party’s position).
9. During the interview, use care not to steer the witness towards a particular version of the facts, whether implicitly or subconsciously, by suggesting answers through questions or otherwise.¹⁷
10. Avoid summarising, to the extent possible the interviewee’s answer, as it may distort the answer. If you are summarising in order to condense the answer, try to use as much as possible the witness’ own words.
11. If the answer is unclear or incomplete, ask the question again or in a different way.¹⁸
12. Avoid needlessly showing the witness other sources of information that may contaminate their memory, such as other witness statements, submissions, etc.
13. Ask whether the witness has contemporaneous notes. If so, ask for a copy of such notes.
14. Have witnesses recount their understanding of the facts before providing them with documents that could affect their memory (e.g. bearing in mind the sheer volume of emails received daily by most senior employees, it is entirely possible that the witness will not have an independent recollection of sending or receiving any given communication, nor of their contemporaneous reaction to it). Use of documents to 'fill in' a chronological narrative may risk altering the recollection of the witnesses as to how the matters at issue developed, if they have no independent recollection of those documents. If a witness is still using such documents to 'fill in' the gaps, the witness should state that they have no independent recollection of those facts and that they are relying on documents.
15. Instruct the witness at the end of the interview:
• to avoid needlessly talking about the testimony with other witnesses; and
• to keep notes of matters that the witness might remember following the interview.
16. If necessary, conduct follow up interviews with the witness in line with the guidelines for the first interview, set out above.

(ii) Assessing information relayed by witness

11. Take into account the period of time elapsed between the experience and its recall.¹⁹
12. Consider to what extent the reliability of a witness’ recollection may be influenced by the effect that facts being recalled could have on the witness’ own position.
13. Are any of the matters at issue likely to result in consequences for the witness personally, in terms of disciplinary proceedings or a risk to a future career progression?
14. Are any of the matters in relation with witness memory a matter of personal embarrassment or sensitivity that may impact how witness memory of the incident was encoded or the way in which it is reported when retrieved?
15. Does the witness have a stake in the importance of its evidence (e.g. does the party’s position on a particular issue depend solely or largely on that evidence)?
16. To what extent has the witness already discussed the issues and/or the evidence with (i) other witnesses, (ii) in-house counsel or management, prior to the witness speaking with external counsel? Has the witness adopted a 'party line' as a result?
17. Consider what extrinsic or neutral contemporaneous evidence supports or conflicts with the witness’ recall as produced in the first meeting with counsel. To what extent are any conflicts material to the issues to be decided by the Tribunal?
18. Consider whether any such conflict should or must be put to the witness before completing preparation of the witness statement, and how to do so without affecting the witness’ recollection.
19. Identify, where possible, any direct conflicts of recall between the witnesses. Is it possible that those conflicts are the result of the way in which the witnesses experienced the relevant event, such that their mental imagery conflicts with the accuracy of the memory (though each witness independently believes in the truth of their recollection)?²⁰
20. Consider the possibility that the witness may have been subject to the 'misinformation effect' of having been told of others’ perspectives (during meetings with other witnesses, senior management, etc.) in a way that modifies the witness’ independent recollection.²¹
21. Explain to the client team, in instances of conflict between witness’ recollection of events, that discussions should be avoided between the client team and the witness which may result in a 'misinformation effect' or a 'confabulation' by the witness.

(iii) Preparation of the witness statement

22. Establish good practice rules at the beginning of the arbitration regarding who should draft the witness statement, the language to be used, the number of drafts, etc. This includes clear communication with the external counsel team, the in-house team, the witnesses themselves, and potentially also the Tribunal.
23. Consider providing the witness with a list of topics/key questions to answer in their own terms/language as a first step, either before or after the initial meeting with the witness.
24. For witnesses capable of doing so, consider having them draft the first draft of the witness statement. Depending on how specific or general the scope of their evidence is, and their facility with the language or experience, the product of that process may need a little, or a lot, of polishing, but it should preserve the witness’ own voice in the giving of their evidence and, at a minimum, ensure that the version of the facts as presented is one with which they are satisfied.
25. Particularly where a witness does not speak the language of the arbitration (or is not fully at ease with the language) steps may be taken to ensure that in the first instance the written testimony is completed in that witness’ first language and translated into the language of the arbitration with care to preserve its idioms.
26. Consider avoiding numerous drafts and re-drafts of the statements. Every iteration of a witness’ evidence following comment or amendment by any other person, including in house or external counsel, is more likely than not to move the witness’ account closer to the pleaded case of the party submitting the testimony.
27. Consider drafting witness statements of co-witnesses independently, i.e. without having witnesses read the drafts of each other’s statements. Although this is common practice in some jurisdictions, the process whereby each witness corroborates the testimony of the others (albeit sometimes with minor deviations) itself runs the risk that the witness’ own recollection becomes co-mingled with and affected by the recollection of the other witnesses.
28. Deploy documents carefully. It is important to clarify whether the witness has any independent recollection of drafting or receiving documents to which the statement may refer. It may even be helpful, from an authenticity perspective, for a witness to explain that they have no independent recollection of a document, but for it to appear as context for the matters which they do independently recall.

(iv) Preparation of the witness ahead of the hearing

29. Consider carefully the extent of witness preparation that is permitted under the rules that apply to the arbitration. In some jurisdictions, it is both permissible and deemed desirable for party counsel to prepare witnesses for the hearing on how to respond to the anticipated topics for questioning. In other jurisdictions, advocates are not permitted to confer with the witnesses regarding their testimony in advance of the hearing and 'preparing' a witness may be a breach of professional conduct rules.
30. Taking into account the techniques described above in the context of witness interviews and the preparation of witness statements, care can be taken to avoid memory contamination in the course of preparing a witness ahead of a hearing.

B. Measures to help identify and weigh the distorting influences that might exist (and take them into account when weighing the witness’ testimony)

31. Disputes are the product of events that unfold over the course of time. The memory of witnesses begins with their first involvement with the transaction or incident that gives rise to the dispute and is then both further formed and affected by subsequent events.
32. Basic practices such as good recordkeeping and contemporaneous notes during the course of a transaction or project (pre-dispute) serve both to provide evidence of what happened at a time when the witness’ memory was fresh and unaffected by information learned subsequently and to help the witness later to recall facts. However, the focus here is on practices that come into play after a dispute has arisen.

1) Steps that can be taken by all players

33. All players in the process, including in-house counsel, outside counsel and arbitrators must evaluate, deal with, and can impact witness memory. The more players understand the memory process, the better positioned they will be to help enhance the value of witness evidence. Therefore, each player can:
1. Educate themselves to understand better the workings of human memory, particularly since research continues to refine the understanding of the workings of human memory.
2. Consider training to be able to conduct cognitive interviews, which are said to yield more accurate information. The conduct of cognitive interviews includes several components, which focus on using basic principles in memory and cognition such as building a rapport with witnesses and having them participate more actively during the interview process, e.g. by reinstating the witness mentally in the situation when the relevant events happened.²²
3. Be aware of what affects memory, including in particular:
• misinformation effect: where (usually misleading) information received after an event interferes with or impairs a witness’ memory of that event;
• memory conformity: where a witness’ memory appears to change in order to match and corroborate potentially conflicting information subsequently provided by another witness;
• age of the witness: elderly individuals may be more susceptible to memory dysfunction;
• vividness of the memory;²³
• one’s own bias affecting one’s perception of witness memory recollection. For example: (i) impact of one’s perception as a result of statements that are well-structured and worded vs less ‘polished’ statements; (ii) coherence and completeness are not necessarily reflective of a better memory; incomplete accounts are the norm in human memory and therefore do not indicate inaccurate memory; extremely coherent narratives may be indicative of prior preparation and rehearsal; (iii) witness confidence is not by itself a good indicator of memory accuracy.²⁴

2) Steps that can be taken by the arbitral tribunal

34. There are steps that an arbitral tribunal may consider taking with a view to reducing the distortions of witness memory and in evaluating witness evidence in the light of distortions. ²⁵
35. For example, where no process has been put in place at the outset, counsel can consider seeking guidance from the tribunal on how to moderate meetings or discussions around the giving of evidence by the witnesses. To date, in the international arbitration context, there is limited guidance on the steps which may be taken by party counsel to 'prepare' a witness, and there are no applicable general standards. The IBA Guidelines on Party Representation in International Arbitration include provisions covering interactions between party counsel and witnesses, which include a permission for the party counsel to assist in the preparation of witness statements and to meet and discuss the witness evidence to prepare witness’ prospective testimony. The comments on these paragraphs of the Guidelines state:

‘The preparation may include a review of the procedures through which testimony will be elicited and preparation of both direct testimony and cross-examination. Such contacts should however not alter the genuineness of the Witness … evidence, which should always reflect the Witness’s own account of relevant facts, events or circumstances ... ‘ ²⁶

36. The tribunal might also consider requiring that each witness statement include information about the way in which it was prepared (particularly if it is not in the witness’ native language) and the extent to which the witness has considered or discussed evidence with the other witnesses. Alternatively, even where such information is not required by the tribunal, counsel for a party may choose to include the information in the written statements of its witnesses with a view to enhancing the reliability of the evidence of those witnesses in the eyes of the tribunal.
37. The tribunal can give instructions to the witness prior to his/her examination at the hearing:
• Alert the witness about the importance of distinguishing between personal knowledge and information gained post-event from secondary sources.
• Explain that it is permissible to answer ‘I don’t know’ or ‘I don’t recall’ as memories are incomplete.²⁷
• Exclude witnesses from the hearing room until they have given their evidence (known in some jurisdictions as 'invoking the rule').
38. As explained at the beginning of this Section, it is not recommended that the steps set out above are taken routinely in the arbitral process. They are steps that, where appropriate, can help to protect the authenticity of witness memory. But whether any particular step is useful, appropriate or even possible needs to be assessed in the specific circumstances of each case. Some of the factors that may be taken into account when making such an assessment are addressed in Section VI ‘Conclusions and Recommendations’ below.

VI. Conclusions and Recommendations

A. Assessing the needs of each specific case

1. Sections III and IV of this Report describe the scientific research that demonstrates how the memory of an honest witness may be at risk of being corrupted by the interactions that commonly take place in the preparation and presentation of that witness evidence in arbitration proceedings.
2. This finding could lead one to conclude that steps should be taken to avoid this potential corruption of witness evidence in all cases. After all, arbitral tribunals may place significant reliance on the evidence of witnesses they consider to be reliable in deciding the case. Parties also invest significant time and cost in the preparation of witness evidence and the witnesses themselves bear what can be a significant personal burden in providing their written evidence and then appearing at a hearing to be cross-examined. Given the significant investment in witness evidence made by all participants in the arbitral process, would it not make sense to systematically take some or all of the steps described in Section V to ensure that the memories of the witnesses who give evidence are as accurate as they can be?
3. In the view of the Task Force, the answer to that question is ‘No’. There are two main reasons for this.
4. First, it is important to take into consideration the different roles that witness evidence can play in arbitration proceedings. As can be seen from the analysis in Section V of this Report, fact witness testimony in arbitration proceedings often does not depend on accurate memory. Where witness evidence is used to provide context or to put a human face to a party’s story (for example), the accuracy of the witness’ memory may not be a relevant issue.
5. Second, even when the accuracy of witness memory is important, it does not necessarily follow that current practices should be changed in order to minimise the risk of memory distortions. There are numerous other considerations that need to be taken into account:
1. One of the key findings of the research summarised in Sections III and IV is that contact between a witness and post-event information, be it in the form of written documents or interaction with other participants, can change memory. Minimising the risk of memory distortions would therefore also mean minimising such contact. In that case, one would have to question practices such as exposing witnesses to written records of the relevant facts, in particular the recollections and statements of other witnesses, etc. Such exposure, however, can also be efficient in terms of refreshing a witness’ memory. Witnesses who testify virtually unprepared are often unhelpful as their testimony remains vague, incoherent and fragmented. A witness who is exposed to useful post event information and whose testimony is thoroughly and professionally prepared by lawyers will be more to the point, consistent and complete.
2. On the other hand, such testimony, usually presented in the form of a witness statement, is often the result of a joint effort between a witness and counsel, which can produce a reconstruction of the pertinent events based on various sources which does not necessarily recount the witness’ actual memory. There is no general answer as to which source is the most accurate. The answer will depend on the specifics of the case and also on the predilections of the participants.
3. Moreover, efforts to keep witness memory pristine must be balanced with practical and/or countervailing considerations. For example, some practices, such as having counsel prepare a witness statement, can have a distorting effect upon the witness’ memory but, nevertheless, have a significant benefit in helping a lay witness communicate to the tribunal the witness’ understanding of the facts. And while one way to avoid memory contamination is to interview witnesses individually, sometimes that is not practical. Even more, sometimes a group meeting is helpful and productive as witnesses can help each other to remember details.
4. While interviewing a witness multiple times with respect to the facts risks altering a witness’ memory and embedding into the witness’ mind post-event information that differs from the reality of what happened, multiple interviews can sometimes allow counsel to distinguish between a witness’ original memory and memories formed by post-event information, and identify what the witness’ personal knowledge really is. Sometimes, the most accurate determination of the facts can be obtained by reviewing with a witness, on more than one occasion, his or her explanation of what happened and presenting the witness with evidence that would appear to be contrary to his or her recollection (in order to allow the witness to explain the seeming inconsistency). While this runs the risk of putting post-event information into the witness’ mind, that process is instinctively logical and natural. It also reduces the risk that at a hearing, the witness, under pressure, will be confronted with contradictory evidence and be unable to explain the seeming contradiction articulately even though the witness, but for being surprised with the evidence, would have a legitimate and rational explanation for the contradiction.
5. Another example which may not always be workable in practice is that the witness drafts the first version of his or her witness statement. Sometimes the witness lacks the sophistication or language skills to draft a well-organised statement that communicates the facts in a logical useful fashion, albeit that the witness is fully capable of accurately and precisely recalling what happened. Alternatively, in the case of some busy high-level executives, their schedules simply do not allow them to take the time to put into writing the knowledge they have and have faithfully communicated to counsel who can prepare the first draft.
6. In essence, the suitability and utility of each of the suggestions in Section VI can be evaluated only in the context of the specific circumstances of each case.

B. The role of the arbitral tribunal

7. Whilst the value of witness evidence is impacted by each of the players in the arbitral process (in-house counsel, outside counsel, party representatives, arbitrators and the witness), the Task Force considers it important to stress certain views regarding the specific role of the arbitral tribunal.
8. Section V above describes various steps that an arbitral tribunal might take with a view to reducing distortions of witness memory and evaluating witness memory in the light of distortions. These include the possibility of inquiring, where useful, into the way in which witness evidence has been prepared in advance of the hearing.
9. The Task Force is firmly of the view that the preparation of witnesses and witness evidence, including the practices identified in Section V, should not become subject to routine inquiry by tribunals or now become yet another topic for requests for documents. Such inquiries should be reserved for exceptional situations, in which witness preparation is an important issue going to the essence of the particular testimony.
10. The importance of delving into witness preparation and witness statement preparation must be measured by the importance of the testimony itself in the context of the dispute and the importance of determining whether a material or important part of the testimony is or is not affected by memory distortion. In any event, any such inquiries must take into account issues of privilege and attorney work product.
11. With these points in mind, when considering steps to take in order to mitigate effects of imperfect memories a tribunal could ask itself the following questions:
1. Should it discuss with the parties the taking of witness evidence at the outset of the proceedings (during the case management conference) to avoid memory contamination?
2. How detailed should such discussion be?
3. Should it issue ‘do’s and don’ts’ in the taking of witness evidence in Procedural Order no. 1?
4. If it provides guidelines at the case management conference, does it put a respondent at a disadvantage considering that claimant has probably already spoken to potential witnesses?
5. Should it require or invite witnesses to include in their witness statements a description of how their evidence has been prepared?
6. Is it useful to investigate at the hearing how witnesses were prepared for the hearing in order to assess the reliability of their evidence?
7. How deep should the investigation by a tribunal be? Should it cover matters such as: how many rounds of interviews, were notes consulted, did the witness review submissions and/or the statements of other witnesses; what preparation did the witnesses undertake for cross-examination?
8. Is such investigation compatible with attorney-client privilege?

C. Avoiding the conclusion that witness evidence is second best

12. The effort to enhance the value of witness evidence takes place in a context of diverse views within the arbitral community. Much of the discussion surrounding the subject of witness evidence is influenced by preconceived notions over the value of witness evidence, the means to present it and to elicit it. Cultural differences and legal traditions must be taken into account. While in some jurisdictions 'witness preparation' is frowned upon or not permitted, in other jurisdictions the failure to prepare a witness would be highly questionable.
13. Driven in part by the fact that memory distortions do affect witness evidence, the suggestion has been made that contemporaneous documents are more reliable than witness evidence. That view is not universally shared. As one Task Force member put it by way of example:

‘The contemporaneous documents are already constructed in order to tell a story; the minutes of meetings are drafted by somebody to create a reality; most people never bother to read them, and therefore they don’t get challenged at the next meeting.’

Contemporaneous documents also come in different forms, some of which will provide more reliable evidence than others; a written transcript of a discussion at a meeting will self-evidently provide more reliable evidence than a contemporaneous note of the meeting made by a meeting participant.

14. As the Task Force believes that witness evidence can be valuable and important, the Task Force considers that a predetermined view of the hierarchy of the value of different types of evidence (such as that documents should be accorded more weight than testimony) is neither justified nor prudent. By enhancing the strength of witness evidence and proposing procedures to enhance the prospects that a witness testimony will reflect the witness’ personal knowledge, the Task Force’s proposals are aimed at ameliorating the concern that in part drives the view that documents should weigh more than testimony. When dealing with something as complex and intangible as how to obtain the maximum value from evidence in general and witness evidence in particular, there is no single 'best' practice. Instead, it is critical to rigorously analyse all the pertinent factors and forces in play that affect the particular witness evidence without preconceived notions as to what evidence is more valuable, witness testimony or documentary evidence, written witness statements versus oral testimony in direct examination, and counsel-led cross-examination versus tribunal-led witness questioning.²⁸
15. While procedures such as cross-examination or tribunal-led questioning and leading questions can (and do) impact the value of witness evidence, the Task Force does not take a position as to which procedure better enhances the value of witness evidence. Accordingly, the suggested procedures do not address such issues; such determinations are best left to the parties and arbitrators in light of the specific circumstances of the case.

D. Training and awareness

16. As noted at the outset, in-house counsel, outside counsel and arbitrators can all take steps to better understand the functioning of human memory and thereby better understand the strengths and weaknesses of the recounting process and know how to aid in that process, in order to get as close to an accurate understanding of what happened as reasonably possible.
17. The Task Force believes that better training of counsel and arbitrators would help to reduce the occurrence of distortions in human memory and reduce the effect of the remaining distortions upon the value of witness evidence. Training courses, with trainers to also include psychologists, and the creation of a database of relevant publications would be valuable resources.
18. In advocating greater awareness of those actions which risk corrupting witness memory, the Task Force is conscious that it is also to some extent providing a road map for those parties and counsel who consider that it would be to their advantage to take those very actions in order to corrupt the memory of their witnesses in a way that supports their case. Not only would such an approach be unethical (and potentially unlawful), but as arbitrators are themselves increasingly alive to the issues, it is undoubtedly highly risky, and very likely to prove counterproductive. Tribunals are typically unimpressed when they get the impression that witnesses are simply ‘towing the party line’. The adverse impression that this creates can not only damage the credibility of the witnesses, but also the credibility of the party itself.
19. Once the process of determining what happened in a dispute is based on a reconstruction of events after the fact, it is destined to be imperfect. The fact that witnesses’ memories are imperfect does not mean that their testimony is not valuable or at times vital. The aspiration should be to reduce the imperfections to the extent reasonably possible in order that the decision rendered can be just, based on a reasonably close approximation of what in fact happened, and in the process enhance the satisfaction of users with the arbitral process.

---

---