Argued April 11, 1934.

OPINION BY MR. JUSTICE MAXEY, June 30, 1934:

The question here is whether or not in an action on an insurance policy, the so-called "presumption against suicide" can take the place of evidence of accidental death in sustaining an averment of death "effected solely through external, violent and accidental means."

On June 3, 1924, Norman C. Watkins obtained a policy of insurance from the appellant company in which the latter agreed to pay the insured's wife, Elizabeth M. Watkins, the sum of $20,000 upon receipt of due proof of the death of the insured during the continuance of the policy. It also agreed to pay the beneficiary "the further sum of $20,000 upon receipt of due proof that such death of the insured occurred during the continuance of said policy while there was no default in the payment of premium, as a result, directly and independently of all other causes, of bodily injuries, effected solely through external, violent and accidental means, of which, except in the case of drowning or of internal injuries revealed by an autopsy, there is a visible contusion or wound on the exterior of the body, and that such death occurred within sixty days of the accident, provided, however, that no accidental death benefit shall be payable if such death resulted from suicide — whether sane or insane; from having engaged in aviation or submarine operations. . . . . . ." On March 19, 1928, Watkins obtained another policy from the same company providing for the payment of sixty monthly installments of $200 each, upon the happening of the same contingencies as those specified in the first policy.

Plaintiff's statement sets forth that on April 16, 1931, "the insured, while engaged in his own pursuits in his garage, without intention on his part, but accidentally, inhaled carbon monoxide gas, as a result of which, directly and independently of all other causes, he died" on the same day. Defendant contested the suit on the ground that the inhalation was not accidental but intentional. The plaintiff claimed $20,000, being the amount of the accidental death benefit under the first policy, and also three installments of $200 each due in April, May and June, 1931, under the second policy, making a total claim of $20,600 with interest. The case was tried before a jury and a verdict was returned for the plaintiff in the sum of $22,582.85. Defendant asked for a new trial. Upon refusal, this appeal followed.

The first assignment of error is that the court charged the jury as follows: "Where it is shown, as it is here, that death resulted from external and violent means, there is a presumption that it was accidental and that it was not intentional." The second assignment of error is that the court charged the jury as follows: "Therefore, the plaintiff's case goes to you necessarily, upon the showing that the death of Mr. Watkins was by external and violent means, and the necessary element of accidental death is, prima facie, supplied by the presumption against suicide." The third assignment of error is based upon an excerpt from the court's charge to the same effect as the foregoing. The fourth assignment of error was the affirmation of plaintiff's first point, reading as follows: "Norman C. Watkins having died a violent death, viz.: by carbon monoxide poisoning, the law presumes that his death was not by suicide but was accidental." The fifth assignment of error was the affirmation of plaintiff's second point, reading as follows: "This presumption has the same probative force and effect as direct evidence of accidental death." The sixth assignment of error was the affirmation of plaintiff's sixth point, reading as follows: "Unless the jury find that the evidence of the defendant outweighs the presumption that Norman C. Watkins did not commit suicide, the verdict must be for the plaintiff." The eighth assignment of error was the refusal of defendant's fourth point, reading as follows: "There is no presumption in this case that Norman C. Watkins died an accidental death."

The court erred in giving to the so-called "presumption against suicide" the weight of a probative fact, thereby casting upon the defendant the burden of proving that the insured's death was something other than the kind of death insured against. Considerable confusion appears in judicial opinions as to the nature of presumptions and their function in the administration of justice. They are not evidence and should not be substituted for evidence. Presumptions are generally grouped into two major classes: (1) of law; and (2) of fact. The former usually have the force of legal maxims and become rules of law, with definite procedural consequences. As Mr. Justice AGNEW said in Tanner v. Hughes and Kincaid, 53 Pa. 289: "A legal presumption is the conclusion of the law itself of the existence of one fact from others in proof, and is binding on the jury, prima facie till disproved, or conclusively, just as the law adopts the one or the other as the effect of proof." Justice AGNEW also refers to the other kind of presumption as merely "a natural probability," i. e., "an inference of fact of the probability." Wigmore on Evidence, 2d edition, volume 5, section 2491, says: "The distinction between presumptions 'of law' and presumptions 'of fact' is in truth the difference between things that are in reality presumptions and things that are not presumptions at all. A presumption is in its characteristic feature a rule of law laid down by the judge, and attaching to one evidentiary fact certain procedural consequences as to the duty of production of other evidence by the opponent. It is based, in policy, upon the probative strength, as a matter of reasoning and inference, of the evidentiary fact; but the presumption is not the fact itself, nor the inference itself, but the legal consequence attached to it. . . . . . . A 'presumption of fact,' in the loose sense, is merely an improper term for the rational potency, or probative value, of the evidentiary fact, regarded as not having this necessary legal consequence. 'They are, in truth, but mere arguments,' and 'depend upon their own natural force and efficacy in generating belief or conviction in the mind.' (Greenleaf, Evidence, section 44.) They have no significance so far as affects the duty of one or the other party to produce evidence, because there is no rule of law attached to them, and the jury may give to them whatever force or weight it thinks best. . . . . . . So long as the law attaches no legal consequences in the way of a duty upon the opponent to come forward with contrary evidence, there is no propriety in applying the term 'presumption' to such facts, however great their probative significance. The employment here of the term 'presumption' is due simply to historical usage, by which 'presumption' was originally a term equivalent, in one sense, to 'inference.' . . . . . . There is in truth but one kind of presumption; and the term 'presumption of fact' should be discarded as useless and confusing. Nevertheless, it must be kept in mind that the peculiar effect of a presumption 'of law' (that is, the real presumption) is merely to invoke a rule of law compelling the jury to reach the conclusion in the absence of evidence to the contrary from the opponent." Wigmore says further in section 2492: "In strictness, there cannot be such a thing as a 'conclusive presumption.' Wherever from one fact another is conclusively presumed, in the sense that the opponent is absolutely precluded from showing by any evidence that the second fact does not exist, the rule really provides that, where the first fact is shown to exist, the second fact's existence is wholly immaterial for the purpose of the proponent's case; and to provide this is to make a rule of substantive law, and not a rule apportioning the burden of persuading as to certain propositions or varying the duty of coming forward with evidence. The term has no place in the principles of evidence (although the history of a 'conclusive presumption' often includes a genuine presumption as its earlier stage), and should be discarded."

It is obvious that what are originally mere inferences may in time become presumptions of law. If, for example, "in the beginning," 100 human beings had been created, and sometime later 51% or more had died, there would then have arisen an inference that all men are mortal. Later in the course of time, with additional data all supporting the same inference, it would become a presumption of law that all men are mortal. Judge BALDWIN of the Supreme Court of Errors of Connecticut has admirably characterized the nature of presumption as follows, in Ward v. Metropolitan Life Ins. Co., 33 A. 902, 904: "The term 'presumption' is used to signify that which may be assumed without proof, or taken for granted. . . . . . . It is asserted as a self-evident result of human reason and experience. In its origin, every presumption is one of fact, and not of law. It may, in course of time, become a presumption of law, and even an indisputable one. Its truth may be so universally accepted as to elevate it to the position of a maxim of jurisprudence. Its convenience, as a rule of decision, may be so generally recognized as to place it in the rank of legal fictions. But, so long as it retains its original character as a presumption of fact, it has simply the force of an argument: 1 Greenleaf, Evidence, section 44; Steph. Dig. Evidence 246."

"Presumptions of fact are at best but mere arguments, and are to be judged by the common and received tests of the truth of propositions and the validity of arguments": Lawhorn v. Carter, 11 Bush. 7. Professor Thayer in his Storrs Lectures, 1896, before the Law School of Yale University, said: "By a loose habit of speech, the presumption is occasionally said to be, itself, evidence, and juries are told to put it on the scale and weigh it. . . . . . . A presumption itself contributes no evidence, and has no probative quality. It is sometimes said that the presumption will tip the scale when the evidence is balanced. But, in truth, nothing tips the scale but evidence, and a presumption — being a legal rule or a legal conclusion — is not evidence. It may represent and spring from certain evidential facts; and these facts may be put in the scale. But that is not putting in the presumption itself. A presumption may be called 'an instrument of proof,' in the sense that it determines from whom evidence shall come, and it may be called something 'in the nature of evidence,' for the same reason; or it may be called a substitute for evidence, and even 'evidence' — in the sense that it counts at the outset, for evidence enough to make a prima facie case. But the moment these conceptions give way to the perfectly distinct notion of evidence proper — i. e., probative matter, which may be a basis of inference, something capable of being weighed in the scales of reason and compared and estimated with other matter of the probative sort — so that we get to treating the presumption of innocence or any other presumption, as being evidence in this its true sense, then we have wandered into the region of shadows and phantoms."

As they deliberated on this case the jurors had the right to bear in mind the fact that to the average human being, life is more attractive than death. They were thus entitled to consider, on an even balance of the evidence as between accidental death and suicide, the probabilities against suicide; but the court below erred in its interpretation and application of the so-called "presumption against suicide." The court gave to the ordinary probability that a human being will not commit suicide the force of a fact in evidence shifting the burden of proof to the defendant. The court charged the jury (as above noted): "The necessary element of accidental death is, prima facie, supplied by the presumption against suicide." "Norman C. Watkins having died a violent death, viz., by carbon monoxide poisoning, the law presumes that his death was not by suicide but was accidental." These instructions made the fact of accidental death such a to-be-expected concomitant of a violent death by carbon monoxide poisoning as to render it unnecessary for the plaintiff, after prima facie proving the latter, to offer further testimony in support of the cause of action sued upon.

It is true that there is a reasonable inference against death being self-inflicted, but this inference does not have a sufficiently broad factual base to warrant its projection to the height of a presumption of law. Presumptions are not fact suppliers; they are guide-posts indicating whence proof must come; for example, in criminal cases, the presumption of innocence mantling an accused is a challenge of proof, to the asserter of guilt. Sometimes the party who has the burden of persuading the jury does not have all through the case the burden of coming forward with affirmative evidence; as for example, in a criminal case if insanity is pleaded, the burden of proving it is on the defendant, who presumably is sane.

Presumptions arise as follows: They are either (1) a procedural expedient, or (2) a rule of proof production based upon the comparative availability of material evidence to the respective parties, or (3) a conclusion firmly based upon the generally known results of wide human experience, or (4) a combination of (1) and (3). The presumption as to the survivorship of husband and wife meeting death in a common disaster is a procedural expedient. It is not based upon extensive data arising from human experience. An unexplained absence for seven years raises the presumption of the death of the absentee upon the expiration of the last day of the period. This also is a procedural expedient — an arbitrary but necessary rule for the solution of problems arising from unexplained absences of human beings. An example of (2) is the rule requiring persons on trial for doing certain acts which are illegal if done without a license to produce evidence that they belong to the class privileged by license. See Com. v. Wenzel, 24 Pa. Super. 467. The following are examples of (3): (a) an envelope properly addressed and stamped will reach the addressee if the latter is alive; (b) a child born during the wedlock of its parents is legitimate; (c) a person who drives across a railroad crossing will show due care. If the driver is killed at such a crossing, the presumption that he showed due care shifts the burden of proof to the party who defends the action on the ground of the victim's want of care. (In this example, the presumption of the victim's due care is merely the converse of the statement that the burden of proof rests on the asserter of the victim's negligence.) A presumption that a debt is paid after a lapse of a definite long period of time is both a procedural expedient (1) and a conclusion based on the results of wide human experience (3).

The so-called "presumption against suicide" is neither a procedural expedient (1) nor a rule rooted in the consideration that one of the litigants has possession of the most available evidence determinative of the issue trying (2), nor is it a conclusion based on the known results of wide human experience (3). It is merely a permissible consideration of the nonprobability of death by suicide.

In judicial trials, as in the other affairs of life, probabilities are always being weighed by those who have to decide questions of fact, but probabilities are not legal presumptions. That a witness who has an interest in the outcome of a lawsuit will "color his testimony" accordingly, is a probability which jurors have a right to consider, but it is not a presumption. Any fabrication of testimony gives rise to a permissible inference by the jury of the lack of credibility of the entire testimony of the fabricator, but this legitimate inference does not amount to a legal presumption that his entire testimony is false. In the instant case the general probability against suicide did not shift the burden of producing evidence upon the party against whom it operated. Here the burden of proving that the death of the insured was caused through "external, violent and accidental means" was squarely placed on the plaintiff. Under the contract of insurance she, in order to make out a cause of action against the defendant for the amount claimed, had to incorporate in her statement these averments. For a commensurate recovery her proof had to sustain them. What the trial judge aptly called "the necessary element of accidental death" was not (as he said it was) "supplied by the presumption against suicide." No presumptions ever take the place of proof. The so-called "presumption against suicide" is merely an inference or argument; it is not what Justice AGNEW, in the opinion above cited, properly designates as a "legal presumption binding on the jury, prima facie, till disproved."

Accidental death, as opposed to self-inflicted death, is no more to be legally presumed than suicide is to be presumed as against homicide. There are thousands of homicides annually, still more suicides and the number of the latter is surpassed by the number of fatal accidents. Yet it might easily happen in some years that the ratios would be reversed. Chief Justice GIBSON said in Burr v. Sim, 4 Wharton 149, 171, that perils of the sea are not to be presumed. "They are such as may or may not occur." He said further: "A natural presumption arises only from a violent probability, because it is a conclusion drawn by experience from the usual current of things; but no violent probability of death arises from a peril, which, though possible, is remote." When the issue is death by accident against death by suicide, the data as to the respective total number of deaths from these two causes is not in the average mind so decisively balanced against the probability of death by suicide as to "harden" the inference against death by suicide into a presumption of law which shifts the burden of proof to the defendant in a suit on an insurance policy against death as a result "of bodily injuries, effected solely through external, violent and accidental means."

An examination of many cases concerning presumptions, particularly "presumptions against suicide," reveals what has been aptly characterized as "a welter of loose language and discordant decisions concerning presumptions." The confusion of thought and expression in the understanding and application of this so-called "presumption against suicide" has arisen from three main causes: (1) the treating of an ordinary permissible deduction as a compellable assumption, i. e., as a presumption, putting on the person on whom it operates the burden of coming forward with opposing evidence, or suffering the penalty of having the so-called presumption accepted as verity; (2) the treating of this permissible deduction as evidence; and (3) the failure to distinguish between suits on insurance policies like the one here sued on, which insure against death as a result "of bodily injuries effected solely through external, violent and accidental means" and suits on those policies which insure against death but which contain a proviso avoiding the policy if the insured dies by his own act. As to (1), this we have already discussed. As to (2), "It is a fallacy to attribute (as do some judges) an artificial probative force to a presumption, increasing for the jury the weight of the facts, even when the opponent has come forward with some evidence to the contrary": Wigmore on Evidence, 2d edition, volume 5, section 2491. ". . . . . . No presumption can be evidence; it is a rule about the duty of producing evidence": Ibid., section 2511. Presumptions are always independent of evidence. Chief Justice SHAW in Webster's Case, 5 Cushing 295, 320, was careful to note this in his charge. There he said: "All the presumptions of law independent of evidence are in favor of innocence." In Agnew v. United States, 165 U.S. 36, counsel for the plaintiff in error (who was the defendant below) asked the trial court to instruct the jury that the presumption of innocence was "to be treated as evidence giving rise to resulting proof to the full extent of its legal efficacy." This instruction was refused. The Supreme Court of the United States, in an opinion by Chief Justice FULLER, said that the instruction was well declined "on the ground of its tendency to mislead." In Lisbon v. Lyman, 49 N.H. 553, the issue was as to the emancipation of a minor, and the jury was instructed that there was a presumption that children under twenty-one are not emancipated, and that the jury was to take this element into account and weigh it in connection with all the testimony. The Supreme Court of New Hampshire held: "The presumption against the freedom of minors was not an element of evidence and could not be weighed as evidence." As to (3), when in a suit on a policy containing a provision avoiding the policy if the insured dies by his own hand, plaintiff makes out a prima facie case, and defendant seeks to avail itself of the substantive defense reserved in the policy that the loss was due to a cause or risk specifically excepted in the policy, the defense becomes an affirmative one and has the burden of proof. See Bowers v. Great Eastern Casualty Co., 260 Pa. 147, 103 A. 536; Carnes v. Iowa Traveling Men's Assn., 76 N.W. 683, and Home Benefit Assn. v. Sargent, 142 U.S. 691.

An instruction that jurors trying the issue presented here, in balancing the evidence for and against the conflicting theories of accidental death and suicide, may keep in mind the general probability against suicide — a probability arising from "the usual current of things" — is proper. But, to charge the jury that after plaintiff proved the insured's death through external and violent means, the balance of the proof required to maintain the cause of action, that is, "the necessary element of accidental death is, prima facie, supplied by the presumption against suicide," is erroneous. Reason leads to this conclusion and authority supports it.

In Grosvenor v. Fidelity C. Co., 102 Neb. 629, 168 N.W. 596, where the plaintiff in an action on an accident policy alleged that the insured lost his life by accidental carbolic acid poisoning, and the defendant's answer contained a general denial and allegation that the insured took his own life by drinking a deadly poison, it was held that the burden was on the plaintiff to show that death was accidental, and that she must do so by evidence of the actual facts, or a situation from which accident is the reasonable and not a reasonable inference or possibility, and that the plaintiff having offered no evidence a directed verdict in her favor was error. The Supreme Court of Nebraska, in its opinion, stated: "Because men love life and fear death, they instinctively avoid obvious danger. This fact, drawn from experience, is the basis of a presumption, relied upon by plaintiff, that when the cause or manner of death is unknown we infer that it was not suicidal. The inference is not based upon a law of nature which is invariable. Men do frequently commit suicide. It is one of a multitude of legitimate inferences, in which we infer the unknown from the known, having greater or less degrees of probability, which we use in reasoning to arrive at the ultimate fact. . . . . . . When, knowing only that one has died from drinking carbolic acid, you say you are in doubt as to cause, and then, bringing into service the presumption against suicidal intent, you finally conclude that the death was accidental, are you not guilty of that error known in logic as petitio principii? Had you not, in reaching your first conclusion, given the theory of accident the benefit of the truth upon which the presumption is founded? Had you assumed as a fact that the deceased contemplated suicide or was indifferent to life, you might not have entertained the doubt. Let us suppose experience has shown that of all the persons who have died from drinking carbolic acid three out of four were cases of suicide; then, would it not be palpably absurd to infer in the given case that the death was not intentional? . . . . . . The presumption against intentional death can no longer prevail as prima facie proof. The burden was and remained upon plaintiff to prove his case."

In Keefer v. Pacific Mut. Life Ins. Co., 201 Pa. 448, 51 A. 366, this court held that when the burden is upon the plaintiff, as it is here, to prove the death of the insured was caused by external, violent and accidental means, the proof could not be based either upon conjecture or presumption. Justice POTTER there said: "There could be no presumption as between disease and accidental means, within the terms of the policy." We say here that there can be no presumption between accidental cause and suicide, within the terms of the policy.

In Merrett v. Preferred Masonic Mut. Acc. Assn. of America (Supreme Court of Michigan), 57 N.W. 169, it was held: "Until there was some evidence tending to show that death resulted from accident, rather than design, or from natural causes, such as apoplexy or heart failure, there was nothing to go to a jury. There was not a prima facie case of accidental death. The burden of proving accidental death is upon a plaintiff."

In Whitlatch v. Fidelity Casualty Co. of New York (Court of Appeals), 149 N.Y. 45, 43 N.E. 405, it was held to be error for the trial judge to refuse to charge in a suit on an accident policy as follows: "While the law does not presume from death alone that it was suicidal, and evidence must be adduced to show that the deceased did take his life, still the burden of proof is upon the plaintiff to show by a preponderance of evidence that it was done by accident, and without intention to take his life."

In Trust Co. v. Siefke, 144 N.Y. 354, 39 N.E. 358, which was a suit upon an accident policy, the Court of Appeals of New York in an opinion by Chief Justice ANDREWS said: "The burden is upon a plaintiff to establish his cause of action when it is in proper form denied by a preponderance of evidence. See Perley v. Perley, 144 Mass. 104, 10 N.E. 726."

In Johns v. Northwestern Mutual Relief Assn., 90 Wis. 332, 41 L.R.A. 587, the Supreme Court of Wisconsin in an opinion by CASSODAY, J., said: "Under the contract of insurance in the case at bar the plaintiff could only recover by showing that the death was the result of, or brought about by, 'some external cause or accident, and not by disease or any voluntary act of the member.' The burden of proving such facts was upon the plaintiff."

In the case of Hill v. Central Accident Ins. Co., 209 Pa. 632, 59 A. 262, this court approved the emphatic statement of the trial court that when plaintiff alleges in a declaration that a "deceased came to his death by violent, external and accidental means, he should be compelled to prove it." It was there held that "all the circumstances surrounding the tragedy" were admissible in evidence and that "there is no reason why circumstantial evidence should not be considered in a civil as well as a criminal case," and that, "when there is evidence that death resulted from violent and external means, and there are conflicting theories as to how the death occurred, it is for the jury to pass upon the facts."

The conclusions that the foregoing lead to are these:

1. The operative facts of the insurance policy sued upon were "external, violent and accidental means" causing the insured's death, and any evidence, whether direct or circumstantial, that tends to prove the operative facts, is admissible.

2. If there are in evidence credited facts or circumstances or both from which the jury may infer legitimately that the insured's death resulted from accidental means, plaintiff is entitled to recover. As to the test of submissibility of evidence to a jury, see Brown v. Schock, 77 Pa. 471, at 479.

3. On plaintiff rests the burden of proving all the operative facts by a fair preponderance of the evidence. An even balancing of the evidence on the issue of death by accidental means or death by suicide denotes that plaintiff fails to sustain her burden of proof and the verdict should be for the defendant.

Causes of action are always set forth affirmatively and if they are to prevail they must be supported either (1) by facts tending to prove directly the cause of action pleaded or (2) by legitimate inferences from circumstances which have met the tests of admissibility. Mere guesses and conjectures cannot be substituted for legal proof.

In the deliberations of the jury there are permissible inferences (sometimes miscalled "presumptions") rooted in general human experience and which have weight when the evidence, respectively, for and against a fact in issue leaves the jury in a "twilight zone" of doubt as to that fact. Such "presumptions may be looked upon as legally recognized phantoms of logic, flitting in the twilight, but disappearing in the sunshine of actual facts": Mackowik v. Kansas City, St. J. C. B. R. Co. (Supreme Court of Missouri), 94 S.W. 256.

Here plaintiff pleaded as her cause of action that the insured "without intention on his part, but accidentally, inhaled carbon monoxide gas, as a result of which he died." She proved that the death was caused by external and violent means, and the trial judge then said that this much being proved, "the law presumes that his [the insured's] death was not by suicide but was accidental" and that "this presumption has the same probative force and effect as direct evidence of accidental death." This was error, for, as already pointed out by us, plaintiff's invoking of a mere "presumption against suicide" (a presumption unsupported by relevant evidence) cannot serve, on the pivotal issue, as an adequate substitute for the proof the well-pleaded cause of action requires.

The assignments of error heretofore specified are sustained. The judgment is reversed with a venire.

Mr. Justice LINN concurred in the result.