Peter Tillers' Home Page

 

The Fabrication of Facts in Investigation and Adjudication

 Peter Tillers

©1995 & ©1998 & ©2007

 **********************************************

A version of this paper was presented at Oxford University in 1995. My thanks go to Adrian Zuckerman for his gracious invitation. I also want to thank Richard Friedman for his helpful comments about this paper.

***********************************************

 

(The above tree, which was inspired – but not validated – by the work of Glenn Shafer, is just for show. I do not discuss it in this paper. Sorry!)

*****

§1. Fabrication of Evidence and Facts in Investigation and Proof.

In the last quarter of a century a diverse group of gifted legal scholars – Lloyd Weinreb, Mirjan Damaška, John Langbein, Steven Landsman, and John Jackson, among others – have published a series of comparative studies of proof in adjudication and litigation. Epistemology has not always been the principal theme of these illuminating studies; instead of discussing the relative accuracy of different methods of proof, some of these comparative studies emphasize other themes, such as the relationship between culture and proof. Nonetheless, all of the studies to which I refer do compare adversarial proof processes with methods of proof that are thought to be significantly less adversarial. These comparative studies – regardless of their authors' specific concerns and objectives – are bound to provoke renewed debate and deliberation, not only about the relative advantages and disadvantages of adversary and “inquisitorial” procedural systems, but also about the relative epistemic merits and demerits of adversary and non-adversary methods of proof; comparisons of adversary and non-adversary methods of proof quite naturally make us wonder whether adversary or non-adversary systems of proof are more likely to “get the facts right.”

John Langbein – a truly original scholar, as well as a provocative one – has taken a strong position on the question of the relative epistemic merits of adversary and non-adversary systems of proof and procedure. Langbein believes that non-adversary systems are far superior to non-adversary systems. Langbein's work takes the evils of partisanship as its starting point. Langbein believes that adversary proof is dangerous and obnoxious because parties and their professional representatives are partisan actors and such partisan and self-interested actors -- particularly lawyers! -- have a strong propensity to alter and debase evidence, inference, and proof in litigation. This premise informs the "ideal" non-adversary system that Langbein describes. [1] Hence, in Langbein's paradigmatic or ideal non-adversary system, lawyers who represent parties (or other self-interested persons or entities) are permitted to play (at most) a peripheral part in the process of proof. Judges, magistrates, or other disinterested public officials are given the primary responsibility for the conduct of proof in part because this limits the opportunities of partisan actors such as lawyers to contaminate or corrupt the evidence that the trier of fact must ultimately consider.

There are many defenders and defenses of the type of adversary system of procedure and proof that Langbein so detests. The adversary paradigm, however, is often defended largely on non-epistemic grounds – for example, on the ground that an adversary system is better equipped than a non-adversary system to impede governmental oppression or on the ground that in an adversary system professional advocates are better able to act as the "friends" of their clients. On occasion, however, proponents of adversary procedure and proof do rely on epistemic considerations. Langbein's paean to German civil procedure provoked a biting response by Ronald J. Allen, Stefan Kock, Kurt Riecherberg and D. Toby Rosen. [2] Allen and his colleagues – “Allen & Co.,” for short – accused Langbein of misrepresenting and exaggerating the virtues of the German procedural system. They also argued that American adversary procedure has some distinct advantages that the German system lacks. Part of the argument made by Allen & Co. has epistemological overtones.

Allen & Co. suggest that the dangers of partisanship in an adversary proof process are overshadowed by the bureaucratic sloth and official indifference that they think are likely to exist in German-style non-adversary systems of proof and procedure. They suggest that an adversary system of proof is more likely than a non-adversary system to get the facts right because partisan professional advocates are more likely than judges, magistrates, or any other "disinterested" government officials to form and investigate a variety of factual possibilities and hypotheses. Disinterested investigating magistrates, they suggest, are likely to develop a kind of tunnel vision about evidence and factual issues. Partisan lawyers in an adversary system, they maintain, are less likely to have that same kind of tunnel vision about evidence and facts because such partisan lawyers primarily want to win cases for their clients whereas professional judges are likely to put a higher premium on wrapping cases up in a neat bundle, as expeditiously as possible. In short, Allen & Co. suggest that an adversary system is more likely to get the facts right because in an adversary system there are institutional incentives for the formation and investigation of a greater variety of factual hypotheses.

The bitter exchange [3] between Langbein and Allen & Co. has two epicenters. One of those epicenters is the phenomenon of evidentiary contamination – the contamination of evidence, inference, and proof by partisan professional advocates. The other epicenter is the role of imagination in the process of investigation and proof – the imagination and formation of factual possibilities and hypotheses by lawyers, judges, or other people charged with the responsibility of gathering evidence and submitting it for evaluation by the trier of fact. The protagonists in the Langbein-Allen exchange deserve our gratitude and our respect for focusing attention on the role of imagination and contamination in adversary methods of investigation and proof. But it is far less clear that we should accord a similar degree of respect to the conclusions of either Langbein or Allen about the question of the relative (epistemic) merits and demerits of adversarial and non-adversarial methods of investigation and proof. In their exchange neither Langbein nor Allen & Co. made any attempt to spell out what they took to be the meaning of matters such as “imagination” and “contamination.” The consequence of this is that the ultimate conclusions that both Langbein and Allen reach are uninteresting and beside the point. A discussion or debate that does not frame the issues in an appropriate manner cannot hope to resolve them in a satisfactory way. This paper does not purport to reach any conclusions about the comparative advantages of adversarial and non-adversarial methods of proof. But it does attempt to pave the way for more enlightening argument about that question by clarifying the nature of matters such as “creativity” and “contamination” in forensic investigation and proof.

The title of this paper speaks of the “fabrication” of facts. The idea that investigation and proof involve “fabrication” is very much at the heart of this paper. This paper was originally presented in an Oxford honors seminar entitled “Reconstructing Reality in the Courtroom.” That felicitous title suggests that the organizers of that seminar (Adrian Zuckerman, Fellow of University College, and Michael Hart, QC, Fellow of All Souls College) believed that forensic proof necessarily involves creativity as well as fidelity. I want my audience to think along similar lines. Hence, I want my audience to keep in mind, first, that evidence ordinarily does not drop in a lawyer's, a court's, or a fact-finder's lap, and that evidence is ordinarily imagined, discovered, acquired, analyzed, and assembled before finds its way into anyone's lap. American lawyers talk about “putting a case together,” “developing” evidence, and “making” a case. But this kind of lawyers' talk – talk about “shaping” a case, and the like – also serves as an apt reminder that there is a potential downside to the creative, constructive, and imaginative aspects of forensic investigation and proof. This kind of talk serves as a reminder that lawyers (and other actors), while acting as midwives to proof, have the ability to contaminate evidence as well as to collect and assemble it. Similarly, the term “fabrication” should serve as a reminder both that facts are artifacts and that partisan participants in the process of proof may try to make something out of nothing.

 

§2. A Slick Cover Story.

Consider the following story:

A

Your name is Able, Able Attorney. Your nickname is Slick. You are a lawyer. You work in Middletown, Middle State, U.S.A.

The date is October 5, 1994. The time is 8:45 a.m. You have just arrived in your office. Just then Dale, your legal secretary, ushers Peter Plaintiff into your office.

Dale explains that Peter has asked to see you. You are irritated; you haven't had a chance to drink your coffee, check your phone for messages, or consider whether you even want to see Peter. Nonetheless, you say, "I'm very glad to meet you, Peter. You don't mind if I call you Peter, do you? How can I help you?"

Peter tells you that he wants to explore the possibility of bringing a personal injury action against the owner of a "convenience" store. You ask Peter for details.

 

B

Peter tells you that he was shot in the head by a clerk at a "Happy Valley" store in Madison City (in Middle State) a few months ago. He explains that he thinks that the clerk who shot him was negligent, reckless, or malevolent.

Peter adds, "I personally lean toward the malevolence hypothesis. I know that clerk. He's a vicious s.o.b. His name is Harry. His nickname is Dirt. And that's no accident."

You ask Peter for further details. He tells you that he went to the Happy Valley store on the night of April 1, 1994. He says that he remembers entering the store at 10:45 p.m. He also vividly remembers seeing a blinding flash immediately after entering the store.

"The next thing I remember," Peter says, "is coming back to consciousness just as two white-coated men were carrying me out of the convenience store on a stretcher."

Peter adds, "They took me to a hospital – General Hospital – where I stayed for about three months. When I got to the hospital that night, the doctors told me I had been shot in the head. They took me right into surgery. Later I had two more operations. I'll probably have several more. I hope you'll keep this in mind. I don't talk or think as well as I once did."

You ask Peter, “What makes you think that the clerk at that store – Harry, was it? – what makes you think that Harry was the guy who shot you? Did you see him shoot you?”

Peter answers, “No, I didn't. But the local police told me that they suspect that Harry did the shooting.” You ask, “Did they interview Harry? Did he confess?”

Peter says, “I don't know for sure but I don't think so. The local newspaper reported that Harry disappeared after the shooting. I guess he fled.”

You ask, “Why would he have shot you?”

Peter states, “I dunno. My guess is that he thought I was a robber and that he plugged me when I walked through the door that night.”

You say, “That's malicious?”

Peter states, “Yeah, I think it was. Harry wasn't the kind of person who worried a lot about other people's welfare. He was kind of paranoid. He probably shot me because I wasn't walking right or because I was squinting my eyes. Who knows? But there's one thing I know for sure: Harry was a bad dude. He was bound to shoot somebody some day. Happy Valley Company had no business hiring him.” 

C

The information that Peter has given you seems pregnant with possibilities, problems, issues, and puzzles. You think to yourself, “Well, I can see how this one might play out. Our hypothesis will be that Peter, an innocent customer, was shot by Harry, either because Harry, not caring for anyone but himself, recklessly leaped to the false conclusion that Peter was about to commit a robbery or simply because Harry, an evil fellow, was in a bad mood and just felt like shooting someone that night. In either case, Harry shot Peter. After he did that, he decided that he had better take off. His reason? Simple. He wanted to avoid taking the rap for shooting Peter. He knew he had done wrong and he knew that if he hung around he was going to go to the slammer for having done wrong.”

You think to yourself, “That – or something like that – is probably going to be our story.” However, you also think to yourself, “But Happy Valley Company, of course, is going to have a different story. It may say that when Peter walked into the Happy Valley store in Madison City on the night of April 1, 1994, Peter was interrupting a robbery in progress. That's why Peter was shot in the head: one of the culprits – or the culprit, as the case may be – shot Peter in the head either because he didn't want any witnesses or because he didn't want anyone to interfere with the ongoing robbery. In either case, Harry didn't shoot Peter; a robber did.”

You then think to yourself, “That's not a bad story, is it? The robbery hypothesis explains Harry's mysterious disappearance. The culprit or culprits probably kidnapped Harry. They probably took him far away, killed him, and disposed of his body in a relatively uninhabited and unfrequented location, which explains why Harry has not been found.”

You also think to yourself, “There is, of course, an altogether different possibility that Happy Valley might pursue: Peter Plaintiff went to the Happy Valley store to kill Harry Dirt, whom he obviously didn't like, and Harry shot Peter in self-defense. Or perhaps Peter had suffered some financial misfortune and he went to the store that night with the idea of robbing it, and he then got shot when he tried to do so. But if either one of these things happened, it's hard to explain why Harry hasn't been found. The Company, however, might theorize that after Harry shot Peter in self-defense he panicked and fled in fright. In any event, these are some of the kinds of stories that the Happy Valley may try to push. I had better get down to work. This case looks promising. I had better make the best of it.” 

D

Slick then addresses Peter, “What do you say to a 60-40 cut? You pay court costs as we go along. For example, you will pay 75 bucks or so for service of process and, if necessary, $1000 or so for a reporter's transcript of a deposition. But otherwise you pay nothing. Nothing, that is, unless we win. If I get an award for you, you take 60% of what I get for you, and I take 40%. I get my cut whether you get the money as a result of a judgment or of a settlement. What do you say?”

Peter responds, “Sounds good to me, Slick. Does this mean you'll represent me?”

Slick states, “Yes. I'll have my secretary draw up the retainer agreement. After you sign it and return it to me, we're in business and I can proceed with the preparation of the case.”

 

§3. Scenarios in the Slick Investigation: An Illustration of the Importance and Role of Conjecture in Investigation:

There was a fair amount of discussion in the Langbein-Allen exchange – predominantly in the Allen camp – about the effect of non-adversarial fact-gathering techniques on the development of factual hypotheses. Although I am not prepared to take sides in that skirmish, Allen & Co. were, I believe, quite right in emphasizing the importance of the phenomenon of hypothesis formation. Allen & Co. either said or intimated that the effect of legal prescriptions governing investigation and proof cannot be fairly or fully determined without considering how such prescriptions affect the formation of hypotheses and questions about facts. The Allen camp was correct about this point.

It is sometimes said that any answer to a bad question is correct. It is also the case, however, that a correct answer to the wrong question is uninformative or useless. To say of a system of proof that it generally gives correct answers to the questions that are asked says little good about that system if it is the case that system of proof makes it impossible or difficult for the appropriate questions to be presented. The quality of proof in adjudication does not depend solely on the quality of inference from evidence. The right questions have to be formulated and presented to the trier of fact.

Slick's story suggests one way in which certain kinds of hypotheses – scenarios – are constructed and how such constructs might influence investigation. In our previous work [4] David Schum and I suggested that scenarios have two different kinds of events as their ingredients. One set of events consists of "benchmark" events. These are events for which there is some significant evidential support. Another class of events in scenarios is hypothetical or conjectural events. These are events for which there is no direct evidential support and that are the result of nothing more than conjecture. Figure 1, "Scenario Grammar," illustrates this distinction.

 

The conjectural character of hypothetical events does not mean that the hypothesized events in a scenario are entirely arbitrary. As Figure 1 shows, David Schum and I sometimes refer to hypothetical or conjectural events in a scenario as “gap fillers.” This choice of language reflects our judgment that scenarios have their own logic. Once the judgment is made that certain events probably happened in a certain way at a certain time, it is sometimes rational, given such judgments, to hypothesize that some other events also happened. This is because (most of us believe) events in space and time can influence later events in space and time. Hence, we may be acting very rationally if we believe that the (likely) occurrence of certain events – benchmark events – makes the occurrence of certain other events – including gap fillers – more likely.

An investigator's purpose in forming scenarios is akin to, but not the same as, a fiction writer's objective in doing so. An investigator's scenario contains fictive or hypothetical events, but an investigator's purpose in making conjectures about events and in forming scenarios is, at least sometimes, to figure out what actually happened. An investigator's interest in figuring out what actually happened is one reason why the distinction between benchmark events and gap-fillers is important. An investigator's judgment that events are or are not benchmark events can influence the investigator's thinking and his strategy. For example, in some cases the evidence favoring a benchmark event may seem persuasive and powerful. In such cases a rational investigative strategy might be to use such evidentially well-anchored events as the scaffolding on which to construct scenarios. This is because an investigator can sometimes rationally make the judgment rationally make the judgment that her best chance of constructing a scenario that accurately depicts reality is to construct scenarios around the events that seem to be most firmly anchored in evidence and that are most likely to be true. The investigator's thought might be that those scenarios that are constructed around well-anchored and very likely benchmark events have a better chance of being accurate scenarios than do scenarios that are built around unlikely events whose occurrence is very only weakly supported by evidence.

Now consider Figure 2, entitled “Scenario Scaffold.” This figure represents events that are benchmark events in Slick's eyes. It is quite possible that Slick thinks that the benchmark events depicted in Figure 2 are very likely true, almost certainly fact.

 

 

It is possible that Slick treated the events in Figure 2 as the foundation for his conjectures and guesses about how things happened on or about the night of April 1, 1994, and his guesses about why things happened the way they did. For example, in view of the story Peter told him, perhaps Slick took it as more or less settled (at least for purposes of further investigation) that Peter had entered the Happy Valley store at 10:45 p.m. on April 1, 1994, that Peter was shot in the store shortly thereafter, that Harry Dirt disappeared shortly after Peter was shot, and that Peter was taken to a hospital the same night, where he received medical treatment for several months thereafter. Perhaps Slick's strategy was to use events, which he regarded as (relatively) well-established, to form conjectures about the other events that occurred that night. Consider Figure 3. If the darkened circles in Figure 3 represent evidentially well-established benchmark events and the empty circles, entirely hypothetical events, the sequence of circular shapes in Figure 3 depicts one of the scenarios that Slick constructed on the basis of the events that he viewed as pretty well-settled.

 

 

 

The scenario shown in Figure 3 – I shall call it Scenario 1 – may be useful for a variety of purposes. For example, many legal professionals will immediately see that a trial lawyer might be able to use the sort of Scenario shown in Figure 3 as the basis for a “story” that she intends to develop and present to a fact-finder at a trial. But it is equally important to recognize that scenarios of the sort depicted in Figure 3 can also be useful in fact investigation. For example, an investigator might use Scenario 1 to make guesses or predictions about the existence of evidence. Thus, Slick might say to himself, “Well, let's suppose that Scenario 1 is the true story of the night of April 1 and its aftermath. What sort of evidence might I expect to find if Scenario 1 is correct?”

Figure 4, which involves “retroductive” reasoning, illustrates the pattern of Slick's reasoning.

 

 

 

There are several reasons why Slick might go to the trouble of deliberating in the fashion depicted in Figure 4. One reason might be Slick's wish to have a firmer basis for making a judgment about the chances that Scenario 1 accurately represents the events involving Peter and Harry.

Suppose that Slick's purpose – perhaps uncharacteristically – is to “get to the bottom of things.” He thinks to himself, “Before I go forward with this case, I want to have a pretty good idea of my own about what actually happened on the night of April 1.”

Slick might then think to himself, “I think that the best way of getting to the bottom of things is to focus on Scenario 1. Although I am not yet entirely satisfied that Scenario 1 is correct, I think Scenario 1 has the best chance of being the true scenario. Scenario 1 suggests the existence of various kinds or items of evidence. My investigation now should focus on a search for that evidence. I should focus on the evidence that should be out there if Scenario 1 is true because if Scenario 1 is the best bet, my chances of forming a hypothesis (scenario) that is solidly supported by the evidence are best in the case of Scenario 1. Considerations of economy, efficiency, and investigative parsimony require that I start with that hypothesis rather than some other hypothesis.”

If Slick thinks this way, he may be making the following judgment:

P(S₁) > P(S₂)

P(S₁) > P(S₃)

………

P(S₁) > P(S_n) 

That is, Slick thinks: 

I, Slick, believe that the probability of Scenario 1 is greater than the probability of any other scenario that I can imagine.

It is also possible, in the alternative, that Slick is making the following judgment:

 

P(S₁) > S P(S₂) v P(S₃) v .... P(S_n)

That is, Slick may think: 

I, Slick, believe that the probability of Scenario 1 is greater than the combined probability of all of the other scenarios that I can imagine.

If Slick makes either of these two judgments, it would not be surprising if he also concludes that his chances of finding evidence supporting a specific hypothesis about the sequence of events relating to the incident at the Happy Valley store on April 1 are better if he looks for the evidence suggested by Scenario 1 than if he looks for the evidence suggested by any of the other scenarios that he has formulated. Slick's judgment on this point may be correct because Scenario 1, he believes, is more probable than any of the other specific scenario that he views as serious possibilities. Hence, Slick may act on this judgment of his about his comparative chances of finding evidence supportive of a scenario: as illustrated by Figure 4, Slick may decide to go after the evidence (that he thinks is) suggested by Scenario 1.

 

§4. Scenarios in the Slick Investigation: An Illustration of the Variability of Rational Conjectures in Investigation.

Although I have argued (here and elsewhere) that the formation of hypotheses and conjectures can play a very important role in fact investigation, I do not know of any device or procedure that guarantees the formation of pertinent and fertile hypotheses or conjectures. In fact, I do not believe that such devices or procedures exist.

Consider the cover story involving Slick. I have described one scenario Slick might have constructed, given the evidence and information available to him. However, Slick might have invented yet other scenarios. The cover story, for example, suggests that an investigator in Slick's position might also have constructed the scenario – Scenario 2 – shown in Figure 5.

 

 

 

In this scenario the investigator conjectures that Peter Plaintiff is not an innocent victim of a shooting but had a sudden, preexisting need for money, perhaps because of a loss of employment or for some other reason, and that Peter went to the Happy Valley store on the night of April 1 with the idea of taking money from the store.

The cover story about Slick suggests a third scenario that an investigator in Slick's position might have invented. Scenario 3 is depicted in Figure 6.

 

 

 

These are three of the scenarios that Slick invented. However, it seems quite obvious that another investigator in Slick's position might not construct exactly the same scenarios that Slick did. Another lawyer or investigator might construct a larger number of scenarios than Slick did; she might construct fewer scenarios than Slick did; and one or more of her scenarios might resemble Slick's and yet be somewhat different.

There are other ways to describe the possible differences between Slick's scenario-conjectures and those of another investigator in Slick's shoes. For example, Slick and another investigator, while agreeing on one or more primary scenarios, might disagree about variants or relatives of the scenarios about which they agree. For example, suppose that both Slick and another investigator in Slick's shoes construct Scenario 2. The cover story suggests that Slick also constructed two close relatives of Scenario 2, Scenarios 2.1 and 2.2, which are shown in Figures 7 and 8. It is clearly possible that another lawyer or investigator in Slick's position might develop and consider Scenario 2, but not Scenario 2.1 or 2.2.

 

 

 

 ********************************************************************************

 

 

 

I have described various ways (but not all possible ways) in which Slick and other investigators might differ about the scenarios they should consider. Such investigators, despite their differences with Slick, might well be "rational." Nothing in the Schum-Tillers heuristic theory suggests that reason or logic requires the formation of a specific set of scenarios in the presence of a given “data set.” The whole of our argument implies exactly the contrary. For example, assume that Slick has constructed just three scenarios, Scenario 1, Scenario 2, and Scenario 3. Now make the (wildly unlikely) assumption that another investigator in Slick's position would imagine and consider exactly the same scenarios. It seems clear that even if this investigator or lawyer makes the same scenario-conjectures that Slick does, neither reason nor logic obliges such an investigator or lawyer to agree with Slick that Scenario 1 is the most likely of the three scenarios that are in play. If so, even if an investigator's motive were the same as Slick's – that is, even his motive were also to “get to the (real) bottom of things” –, it is entirely possible that this other investigator might decide to go after first, not the evidence associated with Scenario 1, but the evidence associated with either Scenario 2 or Scenario 3. For example, consider Figure 9, which depicts the evidence that an investigator in Slick's position might believe is associated with Scenario 2. An investigator, instead of pursuing the evidence portrayed in Figure 4, might decide to look for the evidence depicted in Figure 9.

 

 

 

Now consider Figure 10, which depicts the evidence that an investigator in Slick's position might think is suggested by Scenario 3. It is also possible that investigator would decide to look first for the evidence suggested by Scenario 3, rather than the evidence suggested either by Scenario 1 or Scenario 2.

 

 

 

Of course, if an investigator thinks that Scenario 2.1 or Scenario 2.1, rather than Scenario 1, 2, or 3, is the most likely scenario, this judgment might also influence the investigator's investigative strategy. Here again the investigator might decide to pursue the evidence whose existence he believes is suggested by scenario that he views as most likely, which in this instance is either Scenario 2.1 or Scenario 2.2.

 

 

**************************************************************************

 

 

The preceding examples of variations in scenarios and of the ways that such variations might influence investigative strategies of investigators illustrate that even if investigation does have rational and logical components, “rational” investigative strategy can vary significantly from person to person. Consider, for example, the three primary scenarios that Slick constructed and the use that he made of them. There are at least three ways in which an investigator in Slick's position might have made judgments about scenarios that differ from the judgments that Slick made. First, there is nothing inevitable about the particular scenarios that Slick formed after he interviewed Peter. As I have already noted, another investigator might have formed different scenarios. Even if another investigator or lawyer had constructed scenarios that resemble Slick's, that investigator might have put somewhat different events in some of those scenarios. Second, even if another lawyer or investigator had formed exactly the same scenarios that Slick did, he might not have reached the conclusion that Scenario 1 is the most likely of the three scenarios. Third, another investigator might have reached different conclusions about the kind of evidence that each scenario (if true) was likely to generate.

I believe that similar conclusions hold with respect to scenario formation and evaluation in general and with respect to all kinds of conjectures and hypotheses in investigative discovery – which is to say that I believe that different rational and reasonable investigators can construct very different conjectures about evidentiary and factual matters and that such people can attach very different probabilities to the conjectures they do actually form and entertain.

 

§5. Implications of Interpersonal Variability in the Formation of Conjectures and Hypotheses: Let a Hundred (Discordant?) Flowers Bloom in Investigation and Proof?

Assume that different individuals are likely to develop distinct hypotheses and conjectures about evidentiary and factual issues. If this assumption can be made, there are nevertheless several thorny issues that have to be considered before one is entitled to conclude that adversary methods of investigation and proof are superior because of their alleged tendency to generate a wider variety of hypotheses about evidence.

The first problem is that too much of a good thing is not a good thing. If one prefers an adversarial system because one believes that an adversarial system (because of its reliance on multiple participants in the process of investigation and proof) generates a wider variety of conjectures and hypotheses, one must have some sense of how many conjectures and hypotheses one wants. Too many conjectures, hypotheses, and possibilities are more likely to paralyze or confuse the mind rather than enlighten it. The trouble is that there is no inherent, objective, or Platonic benchmark, other than subjective judgment, for the optimal number of conjectures for epistemic purposes. This difficulty suggests that the debate about the relative merits of adversarial and non-adversarial systems is improperly framed if it is posed in terms of the sheer number of evidentiary conjectures or hypotheses that one system or another is likely to generate. It is therefore possible that John Jackson is on a better track; he emphasizes, not the sheer number of conjectures and hypotheses developed and considered during investigation and proof, but the importance of satisfying the questions and epistemic concerns of the trier of fact.

A related difficulty with the let-a-hundred-flowers-bloom approach is that this approach implicitly rests on the supposition that there is a kind of "marketplace of conjectures" that guarantees, in the main at least, that superior conjectures and hypotheses will win out. This may in fact be the case, but it is unlikely that it is always, or under all circumstances, the case. And an advocate of an adversary system would have to demonstrate that the multiplication of evidentiary hypotheses would generally promote rather than impede the search for the truth.

The underlying premise of the let-a-hundred-flowers-bloom argument might also be questioned: one might question whether evidentiary conjectures are in fact likely to be more numerous in an adversary system than in a non-adversary system. Suppose that the choice to be made is between (i) an investigative process that is dominated by one person and (ii) an investigative process that is controlled by two or more persons. Furthermore, suppose that the master or mistress of the show in the one-(wo)man investigative process is as much interested in disposing of potential disputes as in resolving factual questions accurately. Finally, suppose that the investigative decisions of the man or woman who runs the one-person investigative show are effectively unchecked. If these conditions obtain, one possible consideration favoring a two-person adversary investigative process is the possibility that there is a greater probability that a greater number of relevant hypotheses and conjectures will be formed, considered, and investigated in an adversary investigative process than in a non-adversary investigative process. In the circumstances that I have just described there is one tolerably obvious reason for thinking that adversary investigative processes work better than non-adversary investigative processes: advocates who are retained and paid by parties (and whose compensation might depend on getting a favorable result for their clients) are less likely to be afflicted with the kind of bureaucratic sloth that might stifle the imagination and diligence of "nonpartisan" investigators such as judges and magistrates. This premise might be disputed, however. It might be argued, for example, that partisan advocates are also sometimes afflicted by lack of imagination and that stringent training and high professional standards can induce investigating magistrates or judges to think as imaginatively and work as diligently as possible or, in any case, more creatively or industriously than most partisan lawyers do. Even if this counterattack succeeds, however, another consideration suggests that a greater number of factual hypotheses and conjectures are likely to be explored in an adversary investigative process than in a non-adversary investigative process.

This argument that a greater number of conjectures are likely to be explored in an adversary system need not and should not rest on the theory that two (or more) minds are better than one and that multiple minds are therefore more likely to generate a larger number of conjectures about facts and evidence. There is nothing in the heavens or on earth that says that a non-adversarial investigative or forensic process must be operated by only one mind. {Compare, e.g., the medieval grand assize.} But there is an entirely different sort of consideration that speaks in favor of an adversarial investigative process. This consideration has to do, not with the extent of the investigators' imagination, but with the kinds of incentives that are at work in an adversary system of litigation. If one supposes that an investigating magistrate's or judge's primary motivation or incentive is to “get to the bottom of things,” it is possible that (s)he will lean primarily toward exploring or investigating the conjectures, hypotheses, and possibilities that seem to her to be the most probable of those she has imagined or constructed. This might seem a rational strategy because an investigator (including an investigating magistrate) may think that the most efficient investigative strategy, the one most likely to conserve investigative resources, is to investigate the conjecture that seems most plausible; pursuing the evidence associated with the most plausible scenario, it might be argued, is most likely to lead to the discovery of probative and dispositive evidence. It might be argued, however, that an adversarial process, precisely because it puts a lower premium on the efficiency of investigation, is more likely to result in the investigation of pertinent hypotheses. This is precisely because the advocate's primary motivation (let us suppose) is not to “get to the bottom of things,” but to win. The desire to win will mean that an advocate sometimes will have an incentive to investigate hypotheses and conjectures that even in her mind seem less likely to be true than some other hypotheses and conjectures. This is simply because it may make sense for an advocate to focus her investigation on the hypotheses and conjectures that are most likely to favor her client, not on those hypotheses and conjectures that she thinks are most likely true.

There is much force in this argument. For example, suppose that Slick thinks that Scenario 2 is the most likely of the lot. In particular, he believes:

P(S₁) = .2

P(S₂) = .5

P(S₃) = .1

Even though Slick thinks that the probability of Scenario 2 is greater than the probability of Scenario 1, Slick may decide to focus his investigation on Scenario 1 rather than Scenario 2 because Scenario 1, unlike Scenario 2, favors his client. It is true, of course, that Slick may think that he would have a better chance of finding evidence confirming a hypothesis if he focused on the evidence suggested by Scenario 2. But Slick doesn't want to confirm Scenario 2; he gets no advantage (let us suppose) from finding evidence that does so. Hence, Slick might rationally say,

 

I recognize that Scenario 1 is relatively unlikely – I think there's only a one in five chance that it's correct – and my chances of finding evidence to confirm that scenario are therefore correspondingly low. Nonetheless, I will investigative that scenario.

Slick might reason with himself this way: 

I do have at least a pretty good chance – maybe one in five – of finding evidence to support Scenario 1. That chance – though lower than the chance of finding evidence confirming one unfavorable scenario (Scenario 2) – is good enough to make it worth spending my client's money to see if that 20% chance happens to pan out. Yes, I know: the probability of a payoff – a positive evidentiary outcome – is less than 50-50. But I know of no rule that says that it is always irrational to bet on an outcome whose chances are less than 50-50. The thing that has to be thrown into the blender, of course, is the value of a positive evidentiary outcome as well as the amount Peter has to pay to get a chance of getting that positive outcome. I think the payoff of a positive investigative result is substantial – if I could find the evidence I have a chance of finding, I think it would give Peter his freedom – and the cost of taking a chance at that outcome – that is, the cost of looking for evidence favoring Scenario 1 – is relatively small – that is, the cost of looking for this somewhat unlikely evidence will, I think, be relatively low. So why not?

This version of the Slick story suggests the possibility that in an adversarial system of investigation and proof there are stronger incentives than in a non-adversarial system for the examination of evidentiary hypotheses and conjectures that are relatively less probable than some alternative evidentiary hypotheses and conjectures. If so, perhaps this is a good thing because perhaps a system of proof (in the most comprehensive sense) is more apt to "get to the bottom of things" if it investigates not only the most probable conjectures and hypotheses but also those conjectures that are comparatively less probable but that still have a probability that is great enough to warrant the expenditure of investigative resources.

However, the very statement of this thesis also tends to puncture it.

The difficulty with the thesis that I have in mind is not that the cost of multiplying factual hypotheses and conjectures and investigating them has not yet been thoroughly considered. The difficulty I have in mind is also not the very real possibility that it is not in fact true that the number of factual hypotheses generated and evaluated is likely to be greater in an adversarial system of investigation and adjudication than in a non-adversarial one. [5]

The difficulty I have in mind is the phenomenon of evidentiary contamination by the adversary process – the contamination and degradation of evidence, investigation, and, ultimately, the process of proof itself by the activities of adversaries and their agents. 

 

§6. A Re-Introduction to the Problem of the Degradation of Evidence, Investigation, and Proof as a Result of Partisan Participation in the Evidentiary Process.

The nature of the problem of partisan contamination of evidence in investigation (and proof) seems tolerably clear. The great virtue of evidence is its capacity to be independent of the whims, desires, and prejudices of the observer. We learn from evidence (we suppose) only because we are never quite sure what it will tell us and because we are never quite in a position to make it say only what we want it to say. But on these premises, evidence is epistemically useful if and only if it resists us and is independent of us. By hypothesis, then: if evidence can be shaped to suit our needs and desires, it loses some or all of its epistemic potency.

Let us suppose then that in an adversary investigative system the adversaries are given access to evidence and sources of evidence. Let us also suppose that when adversaries interact with evidence they generally "contaminate" it. Further suppose that one adversary – Slick – now confronts three conjectures, Scenario 1, Scenario 2, and Scenario 3. Recall that only Scenario 1 favors Slick's client, Peter Plaintiff. Now suppose that Slick thinks that Scenario 1 is wildly implausible. He believes: 

P(S₁) = .005

Let us suppose that Slick also believes that the probability of finding evidence can be no greater than the probability of Scenario 1 and that is must be less than the probability of Scenario 1. It follows that Slick believes that the odds of finding evidence that tends to support the scenario showing Peter's innocence are no better than 1:199.

Will Slick investigate Scenario 1?

Even in an adversary system Slick will often decide against doing so. The long odds I have stated, however, assume that Slick has no control over the shape and character of the evidence that his investigation will "uncover." Recall the situation portrayed in Figure 4.

 

 

If Slick can (effectively) fabricate evidence during his investigation (and if his moral scruples do not stand in his way), the strategy of investigating Scenario 1 clearly may become a rational and prudent one. Thus, for example, although Slick does not expect to find the evidence of the sort described in Figure 4 (because he thinks that Scenario 1 is very improbable) –, it is entirely possible that he does expect to be able to create the sort of evidence described in Figure 4. (Slick is thus giving a new twist to the notion of "fact as artifact.")  

Have I now turned the tables on the debate about the comparative merits and demerits of adversarial and non-adversarial, or partisan and nonpartisan, methods and systems of forensic investigation? If you believe that contamination of evidence is practically inevitable if private lawyers like Slick are allowed to deal in private with sources of evidence (such as witnesses), perhaps you now also believe that non-adversarial investigative processes – even if they do not result in perfectly exhaustive and absolutely bi-directional investigations – are always superior to adversary investigative processes. Your theory, perhaps, is that partisan or adversary investigation inevitably (by hypothesis?) destroys the independence of evidence and that such partisan investigative activity therefore inevitably destroys the informational and probative value of such evidence. From this you perhaps draw the conclusion that though a non-adversarial investigative process is in certain respects epistemically imperfect, it is nevertheless better to have such an imperfect investigative system than to have an adversarial and partisan investigative process that tends to contaminate, degrade, or destroy all of the evidence that it encounters.

Much more than this has to be said and weighed if one is to make a fair assessment of the comparative vices and virtues of partisan and nonpartisan methods of fact investigation requires. I created a straw-man. Speaking through that straw-man, I was too quick to praise nonpartisan investigation and condemn partisan investigation; I exaggerated the virtues of the former and the vices of the latter. We need to consider several – perhaps quite a few – other matters before we condemn partisan or adversarial methods of investigation. Among other things, we need to consider the notion of contamination with much more care; we need to make the concept of contamination more precise. I shall do that a bit later. First, however, I want to reiterate and revisit a very important thesis. The thesis that I now want to explore just a little bit more is the thesis that partisan interaction with evidence has the ability to destroy the informational or probative value of such evidence.

I readily acknowledge that in some of my previous discussions in this paper I assumed altogether too freely that every partisan interaction with evidence and sources of evidence obliterates the informational and probative value of such evidence and sources of evidence. This assumption is almost certainly false. Some partisan interactions with evidence and sources of evidence may enhance the informational and inferential value of evidence and sources of evidence. Moreover, even a partisan interaction with evidence wipes out a portion of its inferential significance, it does not follow, of course, that such partisan interaction necessarily always totally destroys the evidence's informational and probative value. This much I readily acknowledge. (I have no wish to exaggerate.) Nonetheless, by making such concessions and qualifications, I do not want anyone to lose sight of a familiar but central insight about interaction between partisan investigators, on the one hand, and evidence and sources of evidence, on the other hand. This insight I am referring to is, simply, the obvious idea that partisan interactions with evidence have the capacity to degrade or destroy evidence and its informational and inferential value.

There is an associated (if less familiar) insight that must also be kept in mind. It is important to bear in mind that when partisan investigation degrades or destroys evidence, something has been lost that can never be fully recaptured or restored. Degradation of evidence by partisans produces an irreversible epistemic loss.

In the discussion that follows I shall use the concept of irreversibility to explore the meaning of evidentiary “contamination.” By doing that I hope to help lay the groundwork for a more orderly way of thinking about the problems and prospects of both adversarial and non-adversarial systems of forensic investigation and proof.

 

§7. Epistemic Harm from Degradation and Destruction of Evidence; Irreversibility of Evidentiary Degradation Caused by Partisan Investigation.

Suppose that lawmakers have managed to construct an adversary system of investigation that has one striking characteristic: there is a perfectly accurate and perfectly complete record of the course of each party's investigation. Let us further suppose that legal rules proscribe any outright destruction of evidence and that these proscriptions are 100% effective. However, the applicable legal rules do not entirely proscribe forms of investigation which, it is acknowledged, do have the capacity (sometimes) to alter the character, or flavor, of evidence. Instead of proscribing such practices, the law provides that any and all investigative steps by the parties may be made known to the trier of fact by the other party. For example, in a criminal case the police are free to investigate some leads without being required to pursue others; that is a choice that is left to the police. However, if a police investigation is partial and one-sided, the partial and one-sided nature of the police investigation can be shown to the trier of fact. Alternatively, for example, some courts have sometimes said that if improperly leading and unduly suggestive questions were used by, for example, the police while questioning a complaining witness (such as a child), that fact can be disclosed to the trier of fact and such disclosure will be enough to ward off the risk of harm as a result of any improper questioning. Question: Is this sort of disclosure to the trier of fact a full and adequate remedy for any contamination of evidence that may result from biased or one-sided investigative tactics?

It is standard practice in America – and I am sure that it is common practice in the U.K. as well – for lawyers to try to discredit an opponent's evidence or witnesses by showing the biased way in which the opponent's evidence was gathered or elicited. For example, in the O.J. Simpson case, the lawyers for the defendant went to great lengths to try to discredit the prosecution's evidence by showing a systematic investigative bias on the part of the police and the prosecution in favor of evidence inculpating O.J. Simpson.

In challenges to allegedly biased police investigative practices, American courts have sometimes suggested or intimated that the opportunity to show the biased nature of the investigative practices is as good as actually having the evidence that an unbiased investigation would have uncovered; American courts have sometimes suggested or said that if the opportunity to show the pattern of police investigation is there, it is unnecessary to promulgate rules that force the police to use unbiased and impartial investigative practices, that there is nothing to be gained by having such rules.

This theory emphasizing the role of disclosure in preventing investigative abuses and distortions might be referred to as the "light of day" theory. Unfortunately, however, the theory does not hold as much water as its occasional proponents sometimes imagine it does. Actually having non-degraded evidence in one's hands is generally (if not always) better than knowing the reasons for the degradation or absence of evidence. It is true that the reasons for the contamination, degradation, absence, or destruction of evidence are themselves information – they are, in that sense, evidence – and that such information or evidence sometimes enables the trier to draw some inferences, or hazard some guesses, about what it (the trier) could or would have learned from the evidence if it (the trier) had been able to see the evidence in its previous incarnation. But knowing the reasons for the degradation or absence of particular items of evidence is not the equivalent of actually having the degraded or missing items of evidence in one's hands in a non-degraded form. Inferences about the character and content of missing or degraded evidence are inferences – which is to say, they necessarily remain, to some degree, uncertain ; the trier must guess what the nature or content of the degraded or missing evidence would have been had it not been contaminated or destroyed by the actions of a partisan investigator. If it is assumed that the destructive or contaminating actions of the partisan investigator made the evidence less informative than it would otherwise have been (which, however, both logically and practically speaking, may not always be the case), it is clear that although the trier can attempt to draw uncertain inferences about the nature of the evidence in its "primeval" state, the trier is necessarily better off if it actually has the evidence in its original and non-degraded condition and thus knows – knows with certainty – its content and character. [6]

But surely I oversimplify. And, to some extent, my discussion in this section merely clouds and confuses the issues. For example, it is not true that every alteration of evidence by partisan investigation is epistemically undesirable. Moreover, it is true that some contamination or destruction of informative evidence does not make much of a difference. Nonetheless, the problem of evidentiary degradation and destruction due to partisan investigative or preparatory behavior is real and the thesis of the irreversibility of evidentiary degradation and destruction is true. So, if one is to understand the true dynamics and effects of evidentiary degradation due to partisan investigation and intervention, it is necessary to think more clearly about the phenomenon. We cannot condemn a system of investigation and proof merely because in that system partisans contaminate and destroy evidence more often than they do in some alternative system of investigation and proof. We need a conceptual framework for arguing more clearly and coherently than I have so far about partisan participation in investigation and proof and the degradation and contamination of evidence that such partisan participation sometimes produces.

 

§8. Evidence, Time, and Change.

The proposition that partisan human interaction with evidence produces irreversible epistemic changes is a logical consequence of generally accepted principles about the workings of the world and the human mind. It is useful to state those more general or universal principles explicitly.

Consider, first, several fundamental ontological principles:

OP-1 All knowable events occur in space and time; that is, every event is a particular state or configuration of the space-time continuum at a particular point or interval of space and time. 

•An "event" is a particular configuration, or state, of the space-time continuum; events consist of, or are, particular states of the space-time continuum.

•The definition of an "event" is limited to knowable events, which are events in the space-time continuum. 

OP-2 The space-time continuum is dynamic, not static; that is, the states of the continuum (including "events) are not stable, but change.

•Dynamic systems are those whose states vary or change over time. 

•Changes occur over time. (Time is nature's way of preventing everything from happening at the same time.) 

OP-3 The space-time continuum is a causal continuum; that is, particular states of the continuum both influence and are influenced by other states of the space-time continuum. 

OP-3.1 All events in space and time -- all states of the space-time continuum – are influenced by other events in space and time – other states of the space-time continuum. 

OP-3.2 All events in space and time influence other events. 

OP-4 Time moves forward, not backward; that is, a given state of the space-time continuum can change into a later state but not into an earlier state. 

• Restatement of OP-4: Events and conditions in the space-time continuum are part of an acyclical temporal process, and the direction of this process is always from earlier events to later events.

• Thus, changes in states of the space-time continuum are not reversible; a later state cannot revert to an earlier state. 

OP-4.1 All events in space and time (all states of the space-time continuum) are influenced by prior events in space and time (by prior states of the space-time continuum). 

See OP-3, OP-3.1 & OP-4 

OP-4.2 All events in space and time influence subsequent events in space and time. 

See OP 3.2 & OP-4 

OP-4.3 Later events in space and time cannot influence earlier events.

See OP-4

 

Now please take note of some general and fundamental principles regarding evidence:

OP-E1 "Evidence" occurs in space and time; that is, evidence is itself an event in space and time.

See OP-E2, OP-E2.1 & OP-E3, below. 

Def./Stipulation: Evidence. Evidence emanates from sources of evidence.

Def: Natural Event. A natural event or condition is an event or state in the space-time continuum.

OP-E2 Sources of evidence consist of temporal or natural events, states, or conditions; sources of evidence consist, that is, of particular states of the time-space continuum. 

OP-E2.1 Sources of evidence consist exclusively of temporal or natural events, states, or conditions.

OP-E3 Evidence is the behavior or character of sources of evidence, which consist of particular natural events, states, or conditions. 

Comment : Modern Anglo-American juridical epistemology is predominantly empiricist. Philosophical considerations offer strong support for empiricist epistemology, as do modern epistemological theories stemming from fields such as neuroscience and probability theory. Some other cultural and legal traditions almost certainly also accept the notion that spatio-temporal events may be evidence and sources of evidence. However, a strict empiricist epistemology asserts or assumes that extra-temporal evidence and sources of evidence either do not exist or are illegitimate. See OP-1. 

OP-E3.1 Evidence changes as sources of evidence change. 

Comment: The behavior or character of a natural phenomenon changes as that natural phenomenon (be it a thing or person) changes.

OP-E4.1 Sources of evidence change over time. 

See OP-E2 & OP-2. 

Comment: The underlying hypothesis is that the events and conditions in the space-time continuum do not remain stable, that the events in the space-time continuum are part of a dynamic process. Hence, events or conditions change over time or disappear. OP-E2 states that sources of evidence are events or conditions in the constantly changing space-time continuum.

 

Now consider the following propositions about the relationship between investigation and evidence:

 C-9.3 Investigation changes {affects, influences} sources of evidence.

See OP-E2 & A-5 & Def. of "evidence"

The gist of the idea here is that investigation itself is an event and that, like other events, investigation can influence later events, including the behavior of sources of evidence. 

C-9.4 Investigation changes {affects, influences} evidence.

See OP-E3 & OP-E3.1  

Ergo : Changes in evidence and in sources of evidence induced by investigation are irreversible. 

See OP-5 (changes in natural states or conditions are irreversible).

See OP-E2 & OP-E2.1 (a source of evidence is a natural state or condition)

See Def. of "evidence" (evidence is a product of a source of evidence; the existence of evidence and the character of evidence are dependent on the existence and character of a source of evidence)

Comment: The thesis of irreversibility can be restated in the following way: Although it is true that we can draw (some) inferences about past events, including past evidence and sources of evidence, this does not change the fact that an item of evidence that was destroyed or degraded by an investigation destroyed no longer exists or no longer exists in an non-degraded form. Hence, investigation necessarily deprives us of the evidence and information that we would have had in the absence of investigation.

 

***

 

The thesis of the irreversibility of evidentiary degradation can also be expressed graphically.

Consider Figure 13:

 

 

 

Let S_o represent the initial state of a source of evidence, which I will suppose is a witness, Sam Source.

Let i represent an "investigative action," such as the asking of a question.

S_b then represents the state of Sam Source after the investigative action has been taken.

The investigative action i taken by the investigator has induced a change in the condition or state of Sam Source.

The bold arc (line; arrow) from S_o to S_b indicates that investigative action i has “caused” Sam Source to assume state S_b at time T-2.

The arrow – the directed arc – from S_o to S_b indicates that Sam Source can never revert to state S_o. The change made in Sam Source by investigative action i can never be undone. Hence, although Sam Source can and will undergo further changes, he will never revert to state S_o.

If Intrepid Investigator had not taken investigative action i – i.e., if not-i – Sam Source (let us assume) would have migrated into condition S_a.

 

Now please consider Figure 14.

 

 

R-1, R-2, R-3 and R-4 represent reports, or assertions, that Sam Source might have made (viewed from the standpoint of an observer at time T-1). Sam Source actually made report R-4 because Sam Source, having been influenced by Intrepid Investigator's investigative action (e.g., question) i, assumed state S_b. If Intrepid had not taken action i, Sam Source, we believe, would have assumed condition S_a. We can conjecture, or guess, and sometimes reasonably so, that Sam Source in that event would have made the report R-1, R-2, or R-3. Moreover, we can even conjecture, sometimes reasonably so, that it is most likely that he would have made (say) Report R-2. Nonetheless, this is (mere) conjecture; it is (mere) inference. It is not the same thing as actually having Sam Source be in state S_a and actually having him make report R-2. If Sam were in state S_a, we would know – or, at least, we would know better – whether we have R-1, R-2, or R-3. But, having allowed Sam Source to go into state S_b, we can never "undo the damage," we can never get Sam Source into state S_a.

We might try to get Sam Source into a condition that is like S_a. But we cannot in fact get Sam Source into state S_a. Moreover, we cannot even be certain that any state we have actually managed to get him into is (very) much like state S_a. We can in fact only be certain that any state that Sam Source assumes after being in state S_b will not be exactly like state S_a. Hence, by observing Sam Source's behavior in a state at a later time (i.e., in a state that branches out after and under state S_b), we cannot be certain on the basis of that observation about the kind of behavior or report Sam Source would have generated if he had been allowed to go into state S_a. Hence, if one wants to know how a source of evidence would act or appear in the absence of investigative contamination, having good grounds (even very good grounds) to believe how that source might have acted or appeared in the absence of investigative contamination is never quite as good as actually knowing or observing the behavior or appearance of the source in the absence of investigative contamination.

 ***

The general thesis of this section can be stated in a slightly different way with a collection of several other graphs.

Consider first Figure 15, which represents the possible states that an observer at time T-1 envisions that a human source (such as Sam Source) might later assume.

 

Now consider Figure 16, which represents the states that a source in question actually assumed as time passed.

 

 

Finally, consider Figure 17, which represents (in addition to the matters depicted in Figures 16 and 17) reflection , or “Nachdenken.” Nachdenken, literally translated, means after-thinking, or thinking-back. Figure 17 represents after-thinking – after the fact, so to speak – about what might have been; in particular, Figure 17 depicts reflection about the states, characteristics, or attributes that a human source such as Sam Source might have assumed or acquired, but didn't. (The diagram, for the sake of simplicity, does not explicitly depict variations in the reports R that a source S might have made in various states. Cf. Figure 14, supra, which does depict reports R by S.)

 

 

 

§9. The Inevitability of Evidentiary Change: Implications for Argument about Methods of Investigation and Proof.

Much of what I have said above about evidentiary contamination and the irreversibility of evidentiary contamination may seem to bolster the case for the superiority of non-adversarial and nonpartisan investigative methods. But the implications of the analytical apparatus I have sketched for the question of the appropriate design or structure of systems of forensic investigation in fact remain unclear. Thus far I have supplied, at most, a grammar for discussing or considering the advantages and disadvantages of adversarial and non-adversarial methods of investigation and proof.

Consider again Figures 13 and 14, above. Please note that if no investigative act is undertaken, the Sam Source does not remain in state S_o, Sam Source's original state. If no investigative act is taken, Sam Source assumes state S_a. There is an excellent reason for portraying Sam Source's state in this fashion. While we may not be certain of how Sam Source will change in the absence of investigative action, we can be certain that he will change in some way. One constant in our changing and uncertain world is the certainty of change.

Sam Source will change if Intrepid Investigator undertakes investigative action and Sam Source will change if Intrepid Investigator does not undertake investigative action. Hence, the only sure thing we can say about contamination at this point is this: If we can have evidence in a form in which it gives more information than in a form in which is gives less (or less reliable) information, all things being equal, we always prefer to have the (source of) evidence in the form in which it yields more information rather than less. However, this proposition, even if true, does not tell us whether we are more likely to have good sources when methods of investigation are adversarial or when they are non-adversarial.

If the question is whether the changes caused by partisan investigation make the evidence such that it yields less information than the evidence would otherwise yield, we do not yet have a basis for answering the last question. It is not enough to say that adversary investigation contaminates evidence and sources of evidence. Even if there is no partisan investigation, there is probably some (other) kind of investigation. This other kind of investigation may also contaminate evidence; investigation will – in any event – change evidence. Moreover, even if there is no investigation of any kind, evidence and sources of evidence will become contaminated– i.e., they will change or degrade This is in the nature of things. Sources of evidence, like other natural phenomena or states, are bound to change.

The question of the comparative advantages and disadvantages of adversarial and non-adversarial investigative processes therefore does not devolve into the question of the comparative advantages of uncontaminated evidence versus contaminated evidence. That would be no contest at all. Properly formulated, the debate concerning adversarial and non-adversarial investigative methods presents the question of which type of contamination is (relatively) worse. Each type of investigative method or strategy – including a strategy of investigative inaction! – contaminates evidence. The only genuine question is which type of contamination is worse and which is better. What I have said here does not answer this question. What I have said here provides, at most, a way of talking and arguing about the question. Furthermore, there are dimensions of the question that my way of talking – the conceptual apparatus described here – does not address. For example, even if we believe that partisan investigators generally do more epistemic harm (all things considered) than nonpartisan investigators, we still need a way of talking, thinking, and arguing about the effects of interactions among partisan investigators. It is possible (though not certain) that the acts of otherwise destructive people, when taken together, produce beneficial (epistemic) results. We need to think more, that is, about investigation and proof as a "distributed" decision making activity. Richard Friedman is one of the very few Evidence scholars who have begun to work along these lines. But he would be the first to concede, I think, that much remains to be done. [7]

We have a long way to go before any of us can make well-supported global judgments about the comparative epistemic advantages and disadvantages of adversarial and non-adversarial methods of investigation and proof.

 

FINIS