Teachers' Manual to Green, Nesson & Murray, Problems, Cases and Materials on Evidence, 3rd Edition.

CHAPTER X: ALLOCATION, INFERENCE, BURDENS, AND PRESUMPTIONS
Go to: Teacher's Manual Introduction Chapter 10 Contents Next Section

Introduction (1019)

The concept of the burden of proof turns out on analysis to be a complex amalgam of lesser burdens. Wigmore made a great theoretical advance on the subject by distinguishing what he called the burden of persuasion from the burden of production. The burden of persuasion was the burden of convincing the factfinder of the ultimate propositions of the lawsuit. The burden of production as the burden at any given time during a lawsuit of coming forward with evidence, the consequence of not coming forward with evidence being that the party having the burden of production would suffer a directed verdict against him. The jury decided whether a party had met his burden of persuasion. The judge decided whether a party had met his burden of production.

The concept of the burden of persuasion was later refined to the concept of the risk of non-persuasion. This refinement responded to the observation that the party with the burden of persuasion need not necessarily be the source of the evidence which the factfimder fords persuasive. His opponent might shoot himself in the foot by introducing the evidence that the factfmder finds persuasive against him. The "risk of nonpersuasiori" formulation covers this possibility, and expresses the idea that the party who bears the risk of nonpersuasion is the party that loses if the factfinder, on the basis of all of the evidence, is not persuaded to whatever the applicable standard of proof happens to be.

Thayer took Wigmore's distinction between the burdens of production and persuasion and applied the distinction to presumptions. He reasoned that presumptions are needed only in situations in which the evidence is suggestive but insufficient as it stands to justify a finding, and that the only function of a presumption is to force the party disfavored by the presumption to produce evidence. From this conception of the function of a presumption it seemed to follow (and Thayer argued) that once the disfavored party had produced evidence, the presumption had served its function, its force was exhausted, and the case should then proceed as if the presumption had never existed. This is "bursting the bubble."

Put in Wigmore's terms, the bursting bubble does not affect the burden of persuasion, which remains at all times on the plaintiff. All that is affected is a burden of production, which the presumption imposes on the disfavored party. If the disfavored party fails to come forward with evidence on the issue which is the subject of the presumption then he loses by directed verdict. If he does come forward with evidence, then the case goes to the factfinder unaffected by the presumption and with the risk of nonpersuasion still on the party which had been favored by the presumption.

The problem with (this version of ) the bursting bubble is that it produces radically different results in cases which should logically come out the same way. If the disfavored party does not come forward with any evidence, then he loses. But suppose he comes forward with enough evidence to meet his burden of production (a decision the judge makes . without judging the credibility of the offered evidence), but the factfimder, in assessing the credibility of the evidence, gives the evidence no credit whatever. Although the evidence is unpersuasive, it was sufficient to eliminate the presumption and leave the case to be decided on the basis of the remaining evidence. That means that the party which had been favored by the presumption has no evidence beyond that which had invoked the presumption, and is now unaided by the presumption, with the result that that party now loses. The two situations differ radically in result, but in evidentiary terms they differ only because the disfavored party is offered evidence which the factfinder thought was worthless.

This discontinuity in the operation of the bursting bubble led to efforts to prevent the presumption from vanishing quite so quickly and completely. One line of thought was to treat the presumption somehow as evidence and thereby give an assist to the inference from the presumption's predicate to the presumption's conclusion. The idea was that the law generated the presumption as a way of recognizing the suggestive force of the facts which constituted the predicate for the presumption. Why not, then, tell the factfinder about the existence of the presumption and allow the factfinder to consider the combined force of the facts and the presumption?

Commentators made mincemeat of this concept of presumptions as evidence. They pointed out that a presumption is a rule of law, not evidence, and that jurors could have no way of assessing the "weight" of the presumption in the course of their deliberations. See Morgan, Instructing the Jury upon Presumptions and Burden of Proof, 47 HARV. L. REV. 59 (1933). The result is that the idea of treating presumptions as evidence is a recurring initial reaction to the conceptual and procedural problems which presumptions pose, but has never been accepted as a satisfactory resolution.

The major critical response to the bursting bubble theory of presumptions was the "shifting burden" theory. Proposed by Morgan, the logic was simple. Bursting bubble presumptions had too little force. Specifically, they could be dispelled by evidence which was unpersuasive. The solution was to shift to the disfavored party not only the burden of production, but also the burden of persuasion. If the disfavored party failed to come forward with any evidence, then he would lose by directed verdict. If the disfavored party failed to come forward with evidence which the factfinder found persuasive, then he would lose by the factfinder's verdict. Thus the anomalous discontinuity created by the bursting bubble would be removed.

The Morgan shifting burden approach to presumptions is coherent in that it does not produce odd discontinuities. Its coherence is achieved, however, by shifting the risk of nonpersuasion, which is very strong medicine. Operation of a shifting burden presumption would often be decisive to the outcome of a case.

The Morgan shifting burden approach achieves coherence by shifting the burden of persuasion to the disfavored party. Its conceptual flaw lies in its assumption that there is only one. In fact, the onerousness of a burden of persuasion can be varied by changing the severity of the standard of proof to which the factfinder must be persuaded. By varying the standard of proof by which the factfinder is to be persuaded, a number of different burdens of persuasion can be created, some of which can be far less onerous than the burden of persuasion initially assigned to the party favored by the presumption. To put this another way, the idea of the burden of persuasion needs to be more fully stated by asking, persuasion of what? The degree to which a burden of persuasion is onerous depends on the proposition on which the disfavored party must be persuasive and on the standard to which he must persuade.

To illustrate this idea, consider the presumption of sanity in criminal cases. Absent the presumption, the prosecution would have to prove beyond reasonable doubt that the defendant was sane at the time he committed the offense. The presumption operates so as to relieve the prosecution of this burden if the defendant offers no proof to rebut it. But what if the defendant offers proof sufficient to support a finding that he was insane? Under a bursting bubble approach the presumption would vanish, leaving the prosecution with the burden of persuading the jury beyond a reasonable doubt that the defendant was sane. The bursting bubble would thus accomplish a radical shift in the procedural posture of the case, and would do so on the basis of evidence which, although theoretically sufficient to base a finding (this is the standard the judge applies without considering the actual credibility of the evidence), might be found by the factfinder to be quite incredible.

Now, suppose that the presumption of sanity is considered a burden shifting presumption. The burden on the prosecution absent the presumption is to prove beyond reasonable doubt that the defendant is sane. If the burden of persuasion is shifted by a Morgan-type presumption to the defendant, then the defendant would be required to prove beyond reasonable doubt that he is insane -- a radical shift in the procedural posture of the two parties. Incredible evidence offered by the defendant would do him no good. He bears the burden of persuasion which can only be displaced with credible evidence. This burden shifting presumption has, thus, eliminated the discontinuity of the bursting bubble, but has done so by shifting a very heavy burden to the defendant. Could a lesser burden have been placed on the defendant which would still have eliminated the discontinuity of the bursting bubble?

Consider what happens if we recognize that we can vary the degree to which the burden of persuasion imposed on the defendant is onerous by varying the standard by which the factfinder is to judge whether the defendant's proof is persuasive; that is, we can vary the proposition on which the defendant must be persuasive. For example, the presumption could impose on the defendant only the burden of persuading the factfinder that there is a reasonable doubt about his sanity. This would be a true burden of persuasion, but a far less onerous one than requiring the defendant to prove insanity beyond a reasonable doubt. Just because the prosecution starts with a burden of persuading beyond reasonable doubt with respect to a given issue, it does not follow that shifting the burden of persuasion must require the defendant to prove the opposite of the given issue beyond a reasonable doubt.

The example is drawn from a criminal context to better illustrate the possibilities for modulating the operation of burden shifting presumptions by varying the standard of proof. See United States v. Dubs, 520 F.2d 250 (1st Cir. 1975) (text at 1073). The same approach is possible in civil cases. Indeed, this is the approach the Supreme Court is taking in Texas Department of Community Affairs v Burdine (text at 1048).

The evolution of FRE 301 recapitulates the conceptual evolution of presumption theory from Wigmore to Morgan, but in reverse. The Advisory Committee initially opted for the Morgan "shifting burden" approach. As explained in the Advisory Committee Note:

The so-called "bursting bubble" theory, under which a presumption vanishes upon the introduction of evidence which would support a finding of the nonexistence of the presumed fact, even though not believed, is rejected as according presumption too "slight and evanescent" an effect. [Citing Morgan]

The House Committee rejected the burden shifting approach and substituted a presumption-as-evidence proposal. The House Report justified the proposal as follows:

With respect to the weight to be given a presumption in a civil case, the Committee agreed with the judgment implicit in the Court's version [the original version] that the so-called "bursting bubble" theory of presumptions, whereby a presumption vanishes upon the appearance of any contradicting evidence by the other party, gives to presumptions too slight an effect. On the other hand, the Committee believed that the Rule proposed by the Court whereby a presumption permanently alters the burden of persuasion, no matter how much contradicting evidence is introduced -- a view shared by only a few courts -- lends too great a force to presumptions. Accordingly, the Committee amended the Rule to adopt an intermediate position under which a presumption does not vanish upon the introduction of contradicting evidence, and does not change the burden of persuasion; instead it is merely deemed sufficient evidence of the fact presumed, to be considered by the jury or other finder of fact.

The Senate Committee then rejected the House proposal as "ill-advised," approving the observation that "Presumptions are not evidence, but ways of dealing with evidence." In place of the House Committee's proposal the Senate Committee proposed the bursting bubble approach. The Conference Committee then adopted the Senate Committee's bursting bubble approach, and did some additional damage by attempting to explain how the proposal would work.

The end result was that FRE 301 is a very obscure and difficult section to work with.

The casebook's strategy for dealing with civil presumptions is fairly straight-forward. First, we clarify the way in which burdens of production and persuasion work, and the functions they serve when no presumptions are involved. For this purpose we look at the theory that guides the allocation of elements in civil cases.

Second, we clarify the nature of the problem that presumptions try to solve. The problem occurs in recurrent situations in which the evidence is suggestive of a conclusion but reasonable hypotheses remain which are inconsistent with the conclusion and which cannot be eliminated either by credibility judgments or by reasonable assessments of the significance of circumstantial evidence.

Third, we set out possible theories of presumptions which are coherent, calling them for ease of reference the "explanation seeker", the "allocation shifter', and the "issue switcher ".

Fourth, we examine the major Supreme Court case dealing with civil presumptions, Texas Department of Community Affairs v Burdine, to see how close its approach comes to any of the coherent approaches, and to see whether its approach could be overlaid on FRE 301 to make the federal rule more coherent.

Finally, we look at the operation of presumptions in a variety of contexts to practice and refine the application of the ideas that have been developed.

For an excellent discussion of the functional nature of presumptions, see Allen, Presumptions in Civil Actions Reconsidered, 66 IOWA L. REV. 843 (1982).

Graphic Representation:

Envision the process of proof on an issue as linear.

Plaintiff I--------------------------------------------------------------------I Defendant

Assume that the issue to be proved is X. If proof of X is part of the plaintiffs case, then this means that the starting assumption is "not X". If the plaintiff does nothing he loses. Thus the plaintiff has the burden of displacing the starting assumption about X, and displaces that assumption by offering proof of X.

0 1 2 3 4 5

Plaintiff I---------I---------I---------I---------I----------I Defendant

not X X

We can represent various postures for the evidence about X as points along the line. If the proof of X is so weak that no reasonable person could be convinced by it, the judge will direct a verdict in favor of the defendant. This is represented by the line segment (0-1). The proof of X may be such that reasonable people could be convinced by it, but the factfinder is instead affirmatively convinced by the evidence of "not X". This is the line segment (1-2). The proof with respect to X may be such that the factfinder is unconvinced either way. This is the line segment (2-3). The proof may be such that the factfinder is affirmatively convinced of X, although reasonable persons could differ. This is the line segment (3-4) . Point 3 thus represents the standard of proof, for example, the point at which the factfmder in a civil case would find proof of X to a preponderance. The proof could be such that the only conclusion reasonable people could reach is X, and therefore the judge will direct a verdict for the plaintiff. This line segment is (4-5).

The various permutations of presumptions can be represented on this diagram:

The Bursting Bubble: The plaintiff offers evidence of the predicate for the presumption. This evidence may be suggestive of X but is not by itself sufficient to allow the plaintiff to reach the jury.

To invoke the presumption the plaintiff must prove the predicate to some standard of proof. There has been much debate over what the appropriate standard is (a hard debate to judge since the whole concept of the bursting bubble is flawed.) For illustration, assume that the standard is that the judge must conclude that the plaintiff has proved the predicate facts to a preponderance. This would be represented by showing the state of the proof of the predicate to be past point 3 (*) and the state of the proof of X to be between (0-1) (*).

0 1 2 3 4 5

Plaintiff I--------I---------I---------I *--------I---------I Defendant

not predicate ---> @

0 1 2 3 4 5

Plaintiff I-------*I---------I---------I---------I---------I Defendant

not X --> @



Invocation of the presumption means that the starting assumption (not-X) now shifts and becomes X. The defendant now has the burden of displacing this assumption, and can do so by producing evidence sufficient to support a finding of not-X.

0 1 2 3 4 5

Plaintiff I--------I---------I---------I *--------I---------I Defendant

not predicate @ predicate

0 1 2 3 4 5

Plaintiff I--------I---------I---------I *--------I---------I Defendant

@ <-- X



If the defendant does not offer evidence sufficient to push past point 4, he loses by directed verdict. If he pushes past point 4, the bubble bursts, the assumption reverts to not-X.

0 1 2 3 4 5

Plaintiff I-------*I---------I---------I---------I---------I Defendant

not X --> @



A further debate now ensues among bursting bubble theorists: does the plaintiff automatically now survive a motion for directed verdict? From the diagram it does not look that way. But if the factfinder is free to evaluate the credibility of the evidence offered by the defendant, then the factfinder might conclude that the defendant's evidence was incredible and therefore that the plaintiffs case was convincing. In other words, the defendant's inability to present credible evidence of not-X might be considered evidence of X. On this approach, the plaintiffs case would survive a motion for directed verdict and would be decided by the factfinder on the merits. All agree that the standard of proof would not have been changed by the operation of the presumption: for the plaintiff to win, the factfinder would have to find X to a preponderance.

The Shifting Burden: This type of presumption mainly differs from the bursting bubble because of what happens when the defendant has rebutted the presumption. Instead of the assumption that X is true now reverting to the original assumption that not-X is true, the assumption that X is true remains, and the defendant will lose unless he succeeds in persuading the factfinder of not-X. Thus, in order to win without the presumption, the plaintiff had to bring the evidence to:

0 1 2 3 4 5

Plaintiff I--------I---------I---------1*--------I---------I Defendant

not X --> @ X.

By reason of the operation of the presumption, the defendant, in order to win, must bring the evidence to:

0 1 2 3 4 5

Plaintiff I--------I--------*1---------I---------I---------I Defendant

not X @ <-- X,

The classic shifting burden presumption gives the favored party all of the inferential territory between points 2 and 3.

Burdine's Modulated Shifting Burden: If standard of proof is left at the original point on the inferential line, even though the assumption with respect to X is shifted, then the presumption operates rationally and does not give greater benefit than intended. Thus, instead of saying that the defendant, in order to win, must prove not-X to a preponderance (past point 2 coming from defendant's side), the appropriate standard is that the defendant, in order to win, must offer proof of not-X which is sufficiently convincing to the factfinder to leave the factfinder at least in a state of indecision about the issue X (that is, the defendant must at least reach pint 3).

Go to: Teacher's Manual Introduction Chapter 10 Contents Next Section