OPINION

REYNOSO, J.

Martin Searle died as a result of a self-inflicted gunshot wound to the head. Under the terms of his life insurance policy his beneficiary, Alice M. Searle, was precluded from recovering the full value of the policy if death resulted from "suicide, whether sane or insane." We address five questions: (1) Should we reconsider the conclusions reached on a prior appeal in this action concerning the meaning and validity of that clause of the policy notwithstanding the rule of law of the case? (2) Is the policy clause valid and unambiguous? (3) Must the insurance carrier, in order to avoid liability on the policy, carry the burden of proving that the insured's act of self-destruction was intentional and not accidental? (4) If suicidal intent is proven, can the beneficiary nevertheless recover by demonstrating that the deceased was unable to understand the physical nature and consequences of his act? (5) Can the beneficiary recover by demonstrating that the deceased was unable to control his conduct and that the act therefore was the result of an irresistible impulse? As will appear below, we answer the first four questions in the affirmative and the fifth question in the negative.

In May 1975 Martin Searle was issued a life insurance policy in the face amount of $50,000 by defendant Allstate Life Insurance Company (Allstate). Plaintiff Alice Searle, Martin's second wife, was the named beneficiary. In accordance with the terms of the contract, Martin paid his monthly premiums. He had fully performed his monetary obligations to Allstate at the time of his death on March 13, 1976. He was then 37 years old.

Following Martin's death, Alice Searle (Searle) filed a claim with Allstate for payment on the policy. In reliance on the clause that excluded coverage for "suicide, whether sane or insane," within two years of the issuance of the policy, Allstate denied liability and refunded the premiums. Searle accepted the offer, though she later retendered the premiums. Allstate, however, refused the retender.

Searle sued Allstate for the full face value of the policy on the life of her husband. The parties agreed that death resulted from an act of self-destruction. Allstate moved for summary judgment citing the suicide exclusion clause. Its motion was granted on the basis of the undisputed testimony that Martin's death occurred 10 months after the issuance of the policy and resulted from an act of self-destruction. Searle appealed.

The Court of Appeal, in Searle v. Allstate Life Ins. Co. (1979) 96 Cal. App.3d 614 [ 158 Cal. Rptr. 5 ] (hereinafter Searle I ), reversed the summary judgment, finding a triable issue of fact as to whether the insured possessed the requisite intent to commit suicide. The court held that suicide definitionally required the taking of one's life both intentionally and voluntarily or deliberately. ( 96 Cal. App.3d at p. 616 .) The court concluded that if Martin was so insane as to be unable to form the requisite intent, his act was not suicide. The court held, therefore, that the phrase "suicide, whether sane or insane" was ambiguous and illogical and construed the ambiguity against Allstate as drafter of the policy. In conclusion, the appellate court stated, "If Martin was sane and intended to take his own life, then he committed suicide, and Allstate is liable for repayment of the premiums only. If, however, Martin was insane when he took his own life, then he did not commit suicide and Allstate is liable for the full value of the policy." ( Ibid. ) [1]

On remand the parties stipulated that "the remaining issue to be resolved concerns the mental capacity of MARTIN SEARLE at the time of his death in accordance with the decision of Searle v. Allstate Life Ins. Co., 96 Cal. App.3d 614 ." The parties also stipulated that the matter would be submitted to the jury on special verdict forms, that Allstate had issued a policy to Martin Searle, that Martin died within two years of the issuance, that his death was caused by a self-inflicted gunshot wound, that Allstate had tendered and refused retender of the premiums paid by Martin, and that the stipulation would not impair the right of any party to object to the contents of any instruction or verdict form.

At the jury trial before the San Diego County Superior Court, plaintiff presented the following facts. In March 1975, Martin retired from the Navy after completing 20 years of service as a radioman and communications instructor. In July 1975, Martin began to experience physical problems. The condition worsened gradually and he began to have problems climbing stairs and grasping and holding onto objects. Concerned about his condition, Martin decided to consult a physician at Balboa Hospital, a naval hospital in San Diego. In November 1975, Dr. Joseph Izzo, a neurosurgeon, performed surgery to stabilize Martin's condition.

After the operation, Martin's physical condition improved, although his left arm was so weakened that he was eventually unable to lift it over his head. In January 1976, and increasingly through February, Martin became depressed about his condition. Around January 1, 1976, Martin became sexually impotent. Martin made no express threats of suicide to his wife, his employer or his friends at this time, although he discussed his concern about his impotence with his wife, telling her that "if I lose that, you can forget me."

On February 13, 1976, Martin saw Dr. Izzo for the last time. In addition to his impotence, Martin complained about sleeplessness, restlessness and weight gain in the face of a poor appetite. Dr. Izzo noted Martin's depression and suggested he see a psychiatrist. Martin was adamantly opposed to this suggestion, but he agreed to take antidepressant drugs and stay in touch with Dr. Izzo.

The final contact between Dr. Izzo and Martin was a telephone conversation on March 12, 1976, the evening before Martin's death. At trial, Dr. Izzo testified that Martin sounded happy and jovial and wanted to go off the medication.

Following the telephone call, Martin went to bed and had a very restless night. The next morning, Martin hugged and kissed his wife affectionately and said, "Honey, you better hang in there, things might get a little worse."

That day, Martin and Searle discussed the insomnia problem and Martin said he thought he should get more medication from Dr. Izzo and that he was going to see a psychiatrist Monday morning. Martin said he had had a nightmare about being in jail for manslaughter. They discussed their vacation plans and Searle testified that Martin was looking forward to that vacation.

After breakfast, Martin and Searle went shopping and purchased approximately one month's supply of food. Martin also bought some trees and other items for the garden. He spent the afternoon gardening and doing other household chores. Later in the afternoon, the couple discussed going out, but decided it was too expensive and planned to stay at home and go out the following weekend instead. Before dinner, Searle got Martin to dance, which she had not been able to do for some time, but Martin stumbled and became bitter. Following dinner, they sent the children two daughters from Martin's first marriage and a neighbor's child who was staying overnight upstairs for bed.

Searle went upstairs to settle a dispute between the children. Martin then came upstairs and said, in reference to the children, "You're not going to have to get on them anymore."

Searle testified that she went into the master bedroom and Martin followed shortly thereafter. He appeared agitated and reached under the mattress, pulled out a gun, pointed it towards the children's bedroom and ordered Searle to get the neighbor's child and get out of the house. He repeated his order and threatened to "take them to hell with me." When Searle asked, "Honey, what's wrong?" Martin said, "I can't take it anymore." [2] He then pointed the gun at his wife; she asked him not to hurt the children. Without responding, Martin raised the gun to his head and pulled the trigger. Searle testified that at the time the gun went off, Martin's eyes widened and his face registered a surprised look.

Plaintiff also presented the opinion testimony of two expert witnesses. Dr. Izzo testified that Martin was in a state of rage, in a psychotic state, at the time he killed himself. He testified that Martin did not intend to kill himself and that Martin did not understand the nature and consequences of his act. He stated that "in our society, and in the context of our society and how we think" suicide is not the act of a sane person. Dr. Alan Davidson, a psychologist, testified that in his opinion, Martin had lost contact with reality and was psychotic. Dr. Davidson added that Martin did not intend to kill himself, because intention is a rationally motivated type of behavior. He further testified that Martin did not understand the nature and consequences of his act.

Dr. Robert Litman, a psychiatrist and leading expert in the field of suicide and suicide prevention, was called to testify by the insurance company. Dr. Litman testified that he saw Martin as a proud, dominant man to whom it was very important to be in control. Dr. Litman believed that suicide presented an option to Martin, because his physical injuries and threatened sense of masculinity were very real problems which he could not control. Dr. Litman saw many clues that indicated Martin intended to commit suicide. Thus, Dr. Litman interpreted Martin's request that he be taken off the medication to mean that Martin had decided to handle things himself, and Dr. Litman described the dream about being in prison for manslaughter as indicating Martin's feeling of guilt about killing himself. Dr. Litman's opinion was that Martin was experiencing normal depression and characterized the death as "almost a typical case of suicide." Finally, Dr. Litman disagreed with Dr. Izzo and stated that few persons who commit suicide are insane. He used the term "insane" to describe the mental state of a person who could be committed to a hospital against his or her will.

At the conclusion of the trial, the court instructed the jury that Searle had the burden of proving that Martin did not intend to take his own life and charged them to return special verdicts on whether it had been proved by the preponderance of evidence that at the time of the shooting Martin did not have the mental capacity (1) "to intend to take his life, including awareness of the nature and consequences of his act of self-destruction" or (2) "to govern his own conduct." [3] The jury found that Martin did have the mental capacity to intend to take his life and to govern his own conduct at the time he shot himself. On the basis of those special findings (there being no general verdict), the court entered judgment for Allstate.

Searle appeals, challenging the allocation of the burden of proof on the questions of intent and insanity. Additionally, she alleges error in the refusal of the trial court to give her requested instruction on the definition of suicide.

This court has never ruled on the construction of the clause limiting the insurer's liability where death results from "suicide, whether sane or insane." [4] The Court of Appeal concluded in Searle I that the clause is ambiguous because a person who "is insane and thus is incapable of forming the intent to commit suicide" cannot do so. It further held that summary judgment was improper inasmuch as Allstate could not avoid liability for the full value of the policy without a determination that Martin Searle "was sane and intended to take his own life." ( 96 Cal. App.3d at p. 616 .) We believe that the Court of Appeal erred in holding the clause ambiguous and in indicating that sanity is a necessary element of suicide. Apart from that error, we further conclude that the jury instructions charging Searle with the burden of proving the insanity of the deceased, constituted reversible error.

(1) The rule of "law of the case" generally precludes multiple appellate review of the same issue in a single case. The doctrine applies to this court even though the previous appeal was before a Court of Appeal. In United Dredging Co. v. Industrial Acc. Com. (1930) 208 Cal. 705 [ 284 P. 922 ], this court stated: "Where a decision upon appeal has been rendered by a District Court of Appeal and the case is returned upon a reversal, and a second appeal comes to this court directly or intermediately, for reasons of policy and convenience, this court generally will not inquire into the merits of said first decision, but will regard it as the law of the case." ( Id. at p. 712; see also Davies v. Krasna (1975) 14 Cal.3d 502, at p. 507, fn. 4 [ 121 Cal. Rptr. 705 , 535 P.2d 1161 , 79 A.L.R.3d 807 ].) As we stated in People v. Shuey (1975) 13 Cal.3d 835, 842 [ 120 Cal. Rptr. 83 , 533 P.2d 211 ], "[t]he doctrine, as the name implies, is exclusively concerned with issues of law and not fact."

(2) As noted by this court, however, the "doctrine of law of the case ... is not inflexible." ( Davies v. Krasna, supra, at p. 507, fn. 5 .) In England v. Hospital of Good Samaritan (1939) 14 Cal.2d 791 [ 97 P.2d 813 ], a third appeal was taken to this court and we departed from the rule of law of the case noting that "a court is not absolutely precluded by the law of the case from reconsidering questions decided upon a former appeal. Procedure and not jurisdiction is involved. Where there are exceptional circumstances, a court which is looking to a just determination of the rights of the parties to the litigation and not merely to rules of practice, may and should decide the case without regard to what has gone before." ( Id., at p. 795.) But "[i]n the absence of exceptional circumstances of hardship and injustice the need for attributing finality to considered judicial determinations compels adherence to the previous decision." ( Gore v. Bingaman (1942) 20 Cal.2d 118, 123 [ 124 P.2d 17 ].) (3) Though we have recognized that the rule will be disregarded when necessary to avoid an "unjust decision" (e.g., DiGenova v. State Board of Education (1962) 57 Cal.2d 167, 179 [ 18 Cal. Rptr. 369 , 367 P.2d 865 ]), that exception must rest on "a manifest misapplication of existing principles resulting in substantial injustice" and not on mere disagreement with the prior appellate determination. ( People v. Shuey, supra, 13 Cal.3d 835, 846 .)

Here, however, there is another reason, apart from the degree of injustice that might result from our now following Searle I, for our not regarding ourselves bound by that decision as the law of the case. (4) The primary purpose served by the law-of-the-case rule is one of judicial economy. Finality is attributed to an initial appellate ruling so as to avoid the further reversal and proceedings on remand that would result if the initial ruling were not adhered to in a later appellate proceeding. ( People v. Shuey, supra, 13 Cal.3d 835, 840-841 ; People v. Durbin (1966) 64 Cal.2d 474, 477 [ 50 Cal. Rptr. 657 , 413 P.2d 433 ]; Gore v. Bingaman, supra, 20 Cal.2d 118, 123 .) That reason for the rule is inoperative when the court hearing the subsequent appeal determines that there should be a reversal on a ground that was not considered on the prior appeal. The fact that reversal is necessary in any event frees us from the compulsion that the rule of law of the case might otherwise impose on us to follow a ruling in the prior appeal that we now perceive to be manifestly erroneous. ( State v. Zimmerman (1977) 175 Mont. 179 [ 573 P.2d 174, 178 ]; State v. Hale (1955) 129 Mont. 449 [ 291 P.2d 229, 235 ]; Barton v. Thompson (1881) 56 Iowa 571 [ 9 N.W. 899 ]; Pennington v. Gillaspie (1910) 66 W. Va. 643 [ 66 S.E. 1009 ].)

As already noted, we shall conclude herein that the trial court below committed reversible error in misallocating the burden of proof in its instructions to the jury. Since Searle I was an appeal from a summary judgment, the Court of Appeal had no occasion to, and did not, rule on the burden of proof. Because we must reverse on that ground, this litigation will not be prolonged or complicated by our providing guidance on the meaning of "suicide, whether sane or insane" even though the guidance diverges from some of the conclusions drawn in Searle I.

The Court of Appeal reasoned that an insane person could not commit suicide and that, therefore, the phrase "suicide, whether sane or insane" was ambiguous. The weight of authority, however, supports a contrary conclusion. The phrase "suicide, sane or insane," or its equivalent, has been used in life and accident insurance policies for nearly 100 years. In Bigelow v. Berkshire Life Insurance Co. (1876) 93 U.S. 284 [ 23 L.Ed. 918 ], plaintiff beneficiary sued upon a life insurance policy which provided that it would be void if the insured died by suicide, "sane or insane." The Supreme Court stated: "The words of this stipulation, `shall die by suicide (sane or insane),' must receive a reasonable construction." ( Id., at p. 286 [ 23 L.Ed. at p. 919 ].)

In specifically addressing the question of the ambiguity of the phrase, the court stated, "Nothing can be clearer than that the words, `sane or insane,' were introduced for the purpose of excepting from the operation of the policy any intended self-destruction, whether the insured was of sound mind or in a state of insanity.... In the popular, as well as the legal, sense, suicide means, as we have seen, the death of a party by his own voluntary act; and this condition, based, as it is, on the construction of this language, informed the holder of the policy, that, if he purposefully destroyed his own life, the company would be relieved from liability. It is unnecessary to discuss the various phases of insanity, in order to determine whether a state of circumstances might not possibly arise which would defeat the condition.... For the purposes of this suit, it is enough to say, that the policy was rendered void, if the insured was conscious of the physical nature of his act, and intended by it to cause his death, although, at the time, he was incapable of judging between right and wrong, and of understanding the moral consequences of what he was doing." ( Id. at p. 287 [ 23 L.Ed. at p. 919 ].)

Several other jurisdictions have held that a policy which does not insure the risk of "suicide, whether sane or insane" means suicide by intentional self-destruction by an insane or sane person. (See, e.g., Johnson v. Metropolitan Life Insurance Company (3d Cir.1968) 404 F.2d 1202, 1204 ; Strasberg v. Equitable Life Assur. Soc. of U.S. (1952) 281 App.Div. 9 [ 117 N.Y.S.2d 236, 240 ] ["self-destruction sane or insane"]; 9 Couch, Insurance (2d ed. 1962) § 40:41, pp. 671-672.)

(5) Thus, we conclude that the phrase "suicide, sane or insane" must be construed to include suicide committed by insane persons. Although the degree of understanding of the physical nature and consequences of the act was in doubt in the case at bench, we conclude the court in Searle I erred in holding the clause ambiguous and in reasoning that insanity necessarily precludes formation of the intent to commit suicide.

A proper interpretation of the clause is that it exempts the insurance company from liability only if the insured, whether sane or insane at the time, committed the act of self-destruction with suicidal intent. If suicidal intent is negated by a determination that the insured did not understand the physical nature and consequences of the act, then the company may be held liable for the full amount of the policy.

(6a) The trial court instructed the jury that Searle had the burden of proving that Martin did not intend to commit suicide. We conclude that this allocation of the burden of proof was erroneous. In order to avoid liability on the life insurance policy, Allstate had the burden of proving suicide. Such burden, we conclude, requires more than proof of the act of self-destruction, the proof must further establish that the act was committed with suicidal intent, i.e., the purposeful or intentional causing of death. [5]

In Beers v. California State Life Ins. Co. (1927) 87 Cal. App. 440 [ 262 P. 380 ], the insured died from the effect of strychnine poisoning. The insured had been issued a life insurance policy which limited recovery to premiums paid if death resulted from "suicide ..., committed while sane or insane." ( Id. at p. 444.) The sole issue submitted to the jury was whether the poison was administered with suicidal intent. The appellate court affirmed the jury verdict for plaintiff beneficiary because "whether the defendant introduced sufficient satisfactory proof ... to sustain [the defense of suicide] ... was a question to be solved by the jury, with whose conclusion we are not privileged, legally, to interfere, unless the evidence of suicide is upon its face obviously such as to compel us to hold that no other inference but that of intentional self-destruction may reasonably be drawn." ( Id. at p. 457.)

The burden of proving intent is properly allocated to the party charged with proving suicide. It is well established that the burden of proving an excepted risk or condition subsequent which negates full liability is on the insurer and not on the beneficiary. In Executive Aviation, Inc. v. National Ins. Underwriters (1971) 16 Cal. App.3d 799, 806 [ 94 Cal. Rptr. 347 ], the burdens of proof on the insured and insurer were explained. "Although the insured had the burden of proving the contract of insurance and its terms, as well as the loss, the burden of bringing itself within any exculpatory clause contained in the policy is on the insurer."

Specifically, in the context of life insurance policies, this court has stated that "one seeking to recover on an insurance policy must aver the loss and show that it occurred by reason of a peril insured against, but he need not aver the performance of conditions subsequent, nor negative prohibited acts, nor deny that the loss occurred from the excepted risks." ( Dennis v. Union Mut. Life Ins. Co., supra, 84 Cal. 570, 572 .) Provisions avoiding liability in case of suicide, sane or insane, are conditions subsequent (9 Couch, Insurance (2d ed. 1962) § 40:39, p. 670) and therefore the burden of proving the exclusion rests on the insurer.

(7a) We reject Allstate's contention that it has satisfied its burden of proof by means of the pretrial stipulation. Allstate argues that plaintiff Searle, by agreeing to the fact that Martin died of a self-inflicted gunshot wound, conceded that insurer established a prima facie case of suicide and all remaining issues, if any, must be proven by plaintiff.

The insurance carrier argues that self-destruction and suicide are synonymous, relying on Dennis v. Union Mut. Life Ins. Co., supra, 84 Cal. 570 . In Dennis , however, the excepted risk in the policy was death by "self-destruction" ( id. at p. 571) and although the court found that the insurer had "committed suicide, or self-destruction, with a pistol, while temporarily insane," ( id. at p. 572) the suggestion that the two terms are equivalent is mere dicta. Additionally, Allstate cites Black's Law Dictionary (4th ed. 1968, at p. 1602) and quotes the definition: "SUICIDE: Self destruction;" The complete definition, however, reads "Self destruction; the deliberate termination of one's existence, while in the possession and enjoyment of his mental faculties." ( Ibid., italics added.)

(8) It is well established that a stipulation in proper form within the authority of the attorneys is binding on the parties. ( Palmer v. City of Long Beach (1948) 33 Cal.2d 134, 141 [ 199 P.2d 952 ]; 1 Witkin, Cal. Procedure (2d ed. 1970) Attorneys, § 125, p. 137.) However, "[s]tipulations must be given a reasonable construction with a view to giving effect to the intent of the parties and `the language used will not be so construed as to give it the effect of an admission of fact obviously intended to be controverted, or the waiver of a right not plainly intended to be relinquished....'" ( Palmer v. City of Long Beach, supra, at p. 144.) (7b) So construing the stipulation of the parties in the instant matter, we conclude that the issue of suicidal intent was meant to be controverted. The stipulation's provision that mental capacity was in issue is consistent with the conclusion that plaintiff did not agree that Allstate had established suicidal intent.

(6b) The burden of proving suicidal intent was on the insurance company and reversal is proper if the erroneous instruction was likely to mislead the jury and thus become a factor in its verdict. ( Henderson v. Harnischfeger Corp. (1974) 12 Cal.3d 663, 670 [ 117 Cal. Rptr. 1 , 527 P.2d 353 ].)

In the instant case, the effect of the trial court's instruction was to impose a burden of proof which plaintiff was not bound to meet. The jury, upon hearing and reading the instruction, might reasonably conclude that it had to be convinced by a preponderance of the evidence that Martin acted unintentionally before favoring plaintiff. Instead, before Allstate could claim exception from full liability, it had the burden of showing an intentional act of self-destruction. Under the trial court's instructions, Allstate carried no burden of proving any of the contested issues. We conclude that this constituted reversible error.

(9) We also conclude that the court erred in the first appeal by concluding that insanity necessarily implies a lack of mental capacity to commit suicide. (10) Mental capacity is relevant to the determination of whether an act of self-destruction was committed by the insured with suicidal intent. If the insured did not understand the physical nature and consequences of the act, whether he was sane or insane, then he did not intentionally kill himself. Although the ultimate burden of proving suicidal intent by a preponderance of the evidence remains on the insurer, the beneficiary should be permitted to present evidence to negate suicidal intent.

Allstate argues that section 522 of the Evidence Code applies. That section places the burden of proving insanity on the party claiming a person is or was insane. However, under the insuring clause now before us the issue of whether the insured was sane or insane does not enter into our consideration. The language of the first appellate decision is misleading in this regard. The court's conclusion in Searle I that "[i]f ... [the insured] was insane when he took his own life, then he did not commit suicide" ( 96 Cal. App.3d at p. 616 ) is overbroad but is qualified by earlier statements which equate insanity with an inability to form the requisite suicidal intent. "If a person is insane and thus is incapable of forming the intent to commit suicide, how can that person commit suicide?" ( Id. ) If "insanity" is given that restricted meaning, the quoted formulation parallels the position we adopt as well as the position of several other jurisdictions.

Allstate further argues that the phrase "suicide, sane or insane" negates any issue of state of mind. Many jurisdictions have directly addressed this issue, and a distinct split of authority has arisen in construing the "suicide, sane or insane" exclusion found in many life insurance policies. (See 1B, Appleman, Insurance Law and Practice (rev. ed. 1981) § 493, at pp. 348-355; 9 Couch, Insurance (2d ed. 1962) § 40:39 et seq., at p. 670; 43 Am.Jur.2d, Insurance, § 539 et seq., at p. 609; Annot., Insurance: Construction of "Sane or Insane" Provision of Suicide Exclusion (1966) 9 A.L.R.3d 1015 .) While the apparent numerical weight of authority supports the position asserted by Allstate (see 9 Couch, Insurance, supra, § 40:42, at p. 673), a sizeable minority has concluded that if the insured "is so utterly insane that he does not realize what he is doing or that the act will cause death, then a recovery may be allowed." (1B Appleman, Insurance, supra, § 493 at p. 351.) Investigation by the court into the deceased's ability to understand the physical nature and consequences of the act of self-destruction is required in those jurisdictions. (See, e.g., Strasberg v. Equitable Life Assur. Soc. of U.S., supra, 117 N.Y.S.2d 236 ; Muzenich v. Grand Carniolian Slovenian Catholic Union (1941) 154 Kan. 537 [ 119 P.2d 504 , 138 A.L.R. 818 ]; New York Life Insurance Co. v. Dean (1928) 226 Ky. 597 [ 11 S.W.2d 417 ].)

As discussed above, we adopt the position that Allstate must carry the ultimate burden of proving, by a preponderance of the evidence, that the act was performed with suicidal intent. Evidence that the insured did not understand the nature and consequences of the act is relevant and admissible to negate proof of intent. This allocation of proof is supported by Evidence Code, section 500 which states that "a party has the burden of proof as to each fact the existence or nonexistence of which is essential to the claim for relief or defense that he is asserting." In the case at bench, Allstate must prove suicide to defeat Searle's claim of liability. Although the beneficiary may introduce evidence indicating the insured did not intend to commit suicide, the ultimate burden of proving the fact essential to a claim of nonliability rests with the insurer.

The trial court instructed the jury that Searle had the burden of proof not only on the issue of suicidal intent but also on Martin's lack of "mental capacity to govern his own conduct at the time he shot himself" (see fn. 3, ante ). Since the jury found against Searle on both issues, it is unclear what effect the trial court would have given to a verdict that Martin lacked capacity to govern his own conduct even though he was able to form a suicidal intent at the time he shot himself. For guidance on retrial, we consider the relevance, if any, of a finding of inability to govern one's conduct. (11) To put it another way, we consider whether suicidal intent can be negated by proof that the decedent killed himself under the compulsion of an irresistible impulse.

We think not. As already explained, insanity or other mental derangement does not negate suicidal intent if the decedent is shown to have performed the self-destructive act with an understanding of its physical nature and consequences. Proof that the act was impelled by an irresistible impulse would merely establish that "self-destruction was the very result intended, albeit by a deranged mind." ( Johnson v. Metropolitan Life Insurance Company, supra, 404 F.2d 1202, 1204 .) Thus, proof of the irresistible impulse would not be inconsistent with a finding of suicide while insane. ( Ibid.; Nielsen v. Provident Life and Acc. Ins. Co. (1979) 100 Idaho 223 [ 596 P.2d 95, 98 ]; Strasberg v. Equitable Life Assur. Soc. (Sup.Ct. 1949) 196 Misc. 387 [ 91 N.Y.S.2d 903 ]; National Life Ins. Co. v. Watson (1922) 194 Ky. 355 [ 239 S.W. 35 , 35 A.L.R. 156 ].)

In conclusion we hold that the clause limiting the insurer's liability if death results from "suicide, sane or insane" is unambiguous. Further, we conclude that the trial court erred in requiring plaintiff to carry the burden of proving that Martin Searle did not intend to commit suicide and did not possess the mental capacity to intend to commit suicide. Although plaintiff was entitled to present evidence that Martin did not have the capacity to form suicidal intent, the burden of proving suicidal intent should have been placed on Allstate.

The judgment is reversed and the cause remanded for proceedings consistent with this opinion.

Kaus, J., Broussard, J., Grodin, J., and Ackerman, J., [*] concurred.

BIRD, C.J., Concurring and Dissenting.

I concur in the majority's allocation of the burden of proof. However, I cannot agree with their conclusion that the beneficiary of an insured who kills himself while unable to comprehend the moral consequences or general nature of his act can be deprived of benefits under his insurance policy. In reaching their conclusion, the majority rely on 19th century contract law that has long since been rendered obsolete by the modern doctrine of adhesion contracts.

It is not disputed that insurance contracts are contracts of adhesion. Hence, any ambiguities must be read against the insurer. ( Silberg v. California Life Ins. Co. (1974) 11 Cal.3d 452, 464 [ 113 Cal. Rptr. 711 , 521 P.2d 1103 ].) "If the insurer uses language which is uncertain any reasonable doubt will be resolved against it; if the doubt relates to extent or fact of coverage, ... the language will be understood in its most inclusive sense, for the benefit of the insured." ( Continental Cas. Co. v. Phoenix Constr. Co. (1956) 46 Cal.2d 423, 437-438 [ 296 P.2d 801 , 57 A.L.R.2d 914 ].) In determining whether a provision is ambiguous, its terms are to be understood in their ordinary and popular sense. (See generally, Civ. Code, § 1644.)

Applying common dictionary definitions, the clause excluding from coverage "suicide, whether sane or insane" is patently ambiguous. The modifying clause "sane or insane" directly contradicts the term "suicide." Webster's defines "suicide" as "the act or an instance of taking one's own life voluntarily and intentionally ...[,] the deliberate and intentional destruction of his own life by a person of years of discretion and of sound mind ...." (Webster's Third New Internat. Dict. (1961) p. 2286, italics added.) [1] "Insane" is defined as "... exhibiting unsoundness or disorder of mind. ..." ( Id., at p. 1167, italics added.)

Hence, according to the common dictionary definitions, it is doubtful whether an "insane" person could commit suicide. [2] A concerned consumer who took the time to read the full terms of the policy might well place his or her emphasis on the "voluntary" and "deliberate" character of suicide, and thus consider the excluded risk to be under his or her full control. [3]

Nevertheless, the majority maintain that the clause is unambiguous. They point out that their construction is supported by a line of cases stretching back over 100 years. However, the very longevity of this authority exposes its weakness.

The leading case is Bigelow v. Berkshire Life Ins. Co. (1876) 93 U.S. 284 [ 23 L.Ed. 918 ]. (See maj. opn., ante, at pp. 436-437.) Unlike the present majority, the Bigelow court recognized that read literally the clause might be subject to differing interpretations. ( 93 U.S. at p. 286 [ 23 L.Ed. at p. 919 ].) However, under the principles of contract law then in force, there was no doctrine of adhesion contracts. (See generally, Comment, Contracts of Adhesion Under California Law (1967) 1 U.S.F.L.Rev. 306, 307.) Hence, the court was under no compulsion to construe ambiguities against the insurance company.

Instead, the Bigelow court did just the opposite. It sought to effectuate the purposes of the insurance company. The court noted that the term "suicide" standing alone did not apply to an insane person who took his own life. ( 93 U.S. at p. 286 [ 23 L.Ed. at p. 919 ].) "But," the court reasoned, "the insurers in this case have gone further, and sought to avoid altogether this class of risks.... Nothing can be clearer than that the words, `sane or insane,' were introduced for the purpose of excepting from the operation of the policy any intended self-destruction, whether the insured was of sound mind or in a state of insanity." ( Id., at pp. 286-287 [ 23 L.Ed. at p. 919 ].) [4]

The court concluded, as do the majority today, that an insured need only comprehend the physical nature of his act to come within the "suicide" exclusion. ( 93 U.S. at p. 287 [ 23 L.Ed. at p. 919 ]; maj. opn., ante, at p. 439.) Bigelow subsequently spawned a long line of cases applying this construction. (See generally, Annot., Insurance: Construction of "Sane or Insane" Provision of Suicide Exclusion (1966) 9 A.L.R.3d 1015 .)

Bigelow and its progeny do not comport with the modern doctrine of adhesion contracts. Their reliance on the intent of the insurance company as the prime indicator of contractual meaning is entirely out of place under the modern rule that ambiguities must be construed against the insurer. (See Silberg v. California Life Ins. Co., supra, 11 Cal.3d at p. 464 .)

As the above quoted dictionary definitions make clear, the clause is not only ambiguous but also self-contradictory when its terms are given their ordinary and popular meaning. (See ante, at p. 442.) The conflict between the term "suicide" and its modifier, "sane or insane," must be resolved against the insurer. Accordingly, the language "sane or insane" should be given no effect.

In a previous appeal of this case, the Court of Appeal characterized the mental capacity required to commit "suicide" as consisting simply of "sanity." ( Searle v. Allstate Life Ins. Co. (1979) 96 Cal. App.3d 614 [ 158 Cal. Rptr. 5 ] (hereafter Searle I ).) I join the majority in rejecting this misleading characterization.

"Sanity" is an amorphous concept that takes on meaning according to the context of its usage. The issue in the present case is not "sanity" in the abstract, but the mental capacity necessary to commit suicide within the meaning of the exclusion. (Compare M'Naughton's Case (1843) 10 Clark & Fin. 200 [for purposes of the insanity defense in criminal law, "insanity" signifies an inability to distinguish right from wrong].) The Searle I court erred in failing to specify the particular degree of sanity required. It was this omission that engendered the error on remand. [5]

The great majority of jurisdictions hold that where "suicide" is excluded from coverage without the "sane or insane" modifier, the insurer is liable unless the insured was able to comprehend not only the physical nature and consequences of his act, but also its moral character and general nature. (See 43 Am.Jur.2d, supra, at p. 607, and cases cited.) This rule accords with the reasonable expectations of the insured, who is likely to believe that "suicide" is a voluntary and deliberate act under his or her full control. (See, e.g., Webster's Third New Internat. Dict., supra, at p. 2286.)

In conclusion, there is an obvious ambiguity in the phrase "suicide, whether sane or insane." Under the common dictionary definitions of "suicide" and "insane," an insane person cannot commit suicide. An insurance consumer is likely to think of "suicide" as a voluntary act under his or her full control. Under modern principles of contract law, the contradiction must be resolved against the insurer. Accordingly, this court should hold that the "sane or insane" modifier is ineffective. Applying the general rule of construction to a "suicide" exclusion clause, the beneficiary should have an opportunity to prove that the insured was incapable of understanding either the physical nature and consequences of his act or its moral character and general nature. (See 43 Am.Jur.2d, supra, at p. 607, and cases cited.)

In support of their contrary construction, the majority rely on a line of cases stretching back over a century. Unfortunately for their rationale, the very longevity of this authority reveals its weakness. The initial and leading cases rested on the firm foundation of 19th century contract law. Since then, that foundation has crumbled under the weight of the modern doctrine of adhesion contracts. (See generally, Tobriner & Grodin, supra, 55 Cal.L.Rev. at pp. 1273, 1276.) Now is not the time for this court to return to 19th century concepts of contract law.

MOSK, J.

I dissent.

Three significant questions are raised by the major issue in this case the meaning and effect of the standard suicide exclusion clause in life insurance policies. They are: (1) whether the clause is ambiguous; (2) what the phrase "suicide, sane or insane" means; and (3) which party bears the burden of proving suicide. The majority correctly answer the first and third questions. Their answer to the second question, however, is wrong: not only is it contrary to the plain meaning of the clause and against the great weight of authority, but it also reintroduces and indeed exacerbates the very difficulty the clause was intended to remove.

On the first question, it cannot seriously be contended that the phrase "suicide, sane or insane" is self-contradictory and thus ambiguous.

To begin with, a moment's reflection reveals it is the popular understanding that both sane and insane persons commit suicide: we commonly call "suicide" not only the self-destruction of such rational figures as Marcus Junius Brutus and Arthur Koestler, but also such deranged individuals as Adolf Hitler and the Reverend Jim Jones. [1]

Second, for over 100 years the courts of this land have been able to understand and hence have enforced exclusion clauses containing the words "suicide, sane or insane," or similar phrases. (See, e.g., Bigelow v. Berkshire Life Ins. Co. (1876) 93 U.S. 284, 286-288 [ 23 L.Ed. 918, 919-920 ]; Nielsen v. Provident Life and Acc. Ins. Co. (1979) 100 Idaho 223 [ 596 P.2d 95, 97-98 ]; Atkinson v. Life Ins. Co. of Virginia (1976) 217 Va. 208 [ 228 S.E.2d 117, 120-121 ]; Johnson v. Metropolitan Life Insurance Company (3d Cir.1968) 404 F.2d 1202, 1204 [applying N.J. law]; Aetna Life Insurance Company v. McLaughlin (Tex. 1964) 380 S.W.2d 101, 102-106 [ 9 A.L.R.3d 1005 ].)

Third, the statutes of 21 states expressly permit the use in life insurance policies of the very phrase in question here. (Ala. Code, § 27-15-24(a)(2)e (1975); Alaska Stat., § 21.45.250(a)(2)(E) (1956); Ariz. Rev. Stat. Ann., § 20-1226 A.5 (1956); Ark. Stat. Ann., § 66-3323(1)(b)(v) (1980); Del. Code Ann., tit. 18, § 2926(a)(2)(e) (1974); Ga. Code Ann., § 56-2507(1)(e) (1981); Idaho Code, § 41-1925(1)(b)(v) (1977); Ky. Rev. Stat., § 304.15-260(1)(b)(5) (1981); Me. Rev. Stat. Ann., tit. 24-A, § 2525 1.B.(5) (1964); Md. Ins. Code Ann., art. 48a, § 410(a)(5) (1979); Miss. Code Ann., § 83-37-13(9) (1984 Cum. Supp.); Nev. Rev. Stat., § 688A.260 1(5) (1981); N.M. Stat. Ann., § 59A-20-25.A(e) (1985); Okla. Stat. Ann., tit. 36, § 4024 A.3 (West 1976); S.D. Codified Laws Ann., § 58-15-45(5) (1978); Tenn. Code Ann., § 56-7-302(5)(A) (1980); Utah Code Ann., § 31-22-15 (1974); Wash. Rev. Code Ann., § 48.23.260(1)(b) (1961); W. Va. Code Ann., § 33-13-25(a)(5) (1982); Wyo. Stat. Ann., § 26-16-119(a)(ii)(E) (1977); see Colo. Rev. Stat., § 10-7-109 (1973) ["suicide ... after the first policy year ... shall not be a defense ... whether said policyholder was sane or insane"].) Four other states permit essentially identical phrases. (La. Rev. Stat. Ann., § 22:170 B.(3) (West Supp. 1985) ["self-destruction while sane or insane"]; Mont. Code Ann., § 33-20-121 (1981) ["suicide, while seriously mentally ill or otherwise"]; Tex. Ins. Code Ann., art. 3.45(3) (Vernon) (1981) ["death of the insured by his own hand while sane or insane"]; Va. Code, § 38.1-437 (1981) ["die by his own act (whether sane or insane)"].) [2]

On the second of the three questions, however, I part company with the majority. The majority conclude that "[a] proper interpretation of the clause is that it exempts the insurance company from liability only if the insured, whether sane or insane at the time, committed the act of self-destruction with suicidal intent." ( Ante at p. 436.) The majority thus make intent in the technical sense used in the criminal law material to the determination of the trier of fact, and thereby throw open the issue of the insured's state of mind at the moment of the act of self-destruction. In my view, intent in this sense is simply not material; rather, the trier of fact need determine only whether the act of self-destruction, objectively considered, was accidental or, in the popular sense, "unintentional."

As will appear, the plain meaning of the clause excluding "suicide, sane or insane" is that all nonaccidental self-destruction, no matter what the insured's state of mind, is outside the coverage of the policy.

To begin with, such a construction is rooted in the popular understanding, referred to above, that both sane and insane persons commit suicide.

It is consonant, moreover, with the manifest purpose of the clause. By excluding all nonaccidental self-destruction regardless of the insured's state of mind for two years or some similar period of time, the clause promotes two competing goals: broad coverage and low premiums.

The clause increases coverage by excluding suicide for only a relatively short period. In other words, the insurer agrees to "regard suicide after that length of time as one of the hazards covered by the policy." ( Longenberger v. Prudential Ins. Co. of America (1936) 121 Pa. Super. 225 [ 183 A. 422, 425 ].)

The clause reduces premiums by avoiding the costs of suicide in three respects. (See, e.g., Bigelow v. Berkshire Life Ins. Co., supra, 93 U.S. 284, 287 [ 23 L.Ed. 918, 919 ]; Blunt v. Fidelity and Casualty Co. (1904) 145 Cal. 268, 271 [ 78 P. 729 ]; Nielsen v. Provident Life and Acc. Ins. Co., supra, 596 P.2d 95, 97 ; Johnson v. Metropolitan Life Insurance Company (D.N.J. 1967) 273 F. Supp. 589, 594-595 , affd. (3d Cir.1968) 404 F.2d 1202 ; Longenberger v. Prudential Ins. Co. of America, supra, 183 A. 422, 425-426 ; Silliman v. International Life Ins. Co. (1915) 131 Tenn. 303 [ 174 S.W. 1131 , 1133]; Tritschler v. Keystone Mut. Ben. Ass'n. (1897) 180 Pa. 205 [ 36 A. 734, 734-735 ].)

First, the clause obviously avoids the primary costs of suicide that is, the payment of benefits to the suicide's beneficiary. (See Blunt v. Fidelity and Casualty Co., supra, 145 Cal. 268, 271 ; Tritschler v. Keystone Mut. Ben. Ass'n., supra, 36 A. 734, 734 .) As a result, the insurer receives the benefit of the clause by being exempted from liability, and "[t]he assured receive[s] the benefit ... in the reduced amount of the premium...." ( Blunt v. Fidelity and Casualty Co., supra, at p. 271.)

Second, and more specifically, the clause also avoids the costs of fraud the costs of suicide planned, for example, by a terminally ill insured to increase his estate. (See Longenberger v. Prudential Ins. Co. of America, supra, 183 A. 422, 425-426 .) Such costs are not among those that are properly borne by either the general body of policyholders or the insurer. Because insurance is intended and designed to protect against unforeseeable events, [3] the payment of benefits for a planned i.e., foreseeable suicide is antithetical to the very purpose of the contract. Moreover, even if they were proper, the costs of planned suicide are difficult to project and hence difficult to spread in an equitable manner: "While it is possible to compute actuarially the percentage of loss from the ordinary factors of mortality, this becomes somewhat more difficult where an insured voluntarily brings about the maturing event." (1B Appleman, Insurance Law and Practice (rev. ed. 1981) § 492, p. 346.) Such costs would therefore result in either lower profits for the insurer or, more likely, higher premiums for the policyholders. [4]

Third, and of greatest relevance here, a suicide exclusion clause avoids the derivative costs that all suicides, planned or unplanned, would otherwise impose i.e., the large expense in time and money of trying to determine the insured's actual state of mind. (See Johnson v. Metropolitan Life Insurance Company, supra, 273 F. Supp. 589, 594-595 ; Longenberger v. Prudential Ins. Co. of America, supra, 183 A. 422, 426 ; Silliman v. International Life Ins. Co., supra, 174 S.W. 1131 , 1133.) Without an effective exclusion for suicide whether "sane or insane," it would be necessary, in every case in which the beneficiary raised the issue, for the trier of fact to attempt to answer such troublesome questions as whether the insured acted with fraudulent intent (see Silliman v. International Life Ins. Co., supra, at p. 1133) or whether he was a responsible moral agent at the time of the act of self-destruction (see Bigelow v. Berkshire Life Ins. Co., supra, 93 U.S. 284, 286 [ 23 L.Ed. 918, 919 ]; Johnson v. Metropolitan Life Ins. Co., supra, at pp. 594-595). [5]

Because the foregoing questions are difficult and in many cases practically impossible to answer, the derivative costs of suicide are likely to be high. [6] This fact of economic life was recognized over a century ago (see Bigelow v. Berkshire Life Ins. Co., supra, 93 U.S. at pp. 286-288) and is recognized today (see Johnson v. Metropolitan Life Insurance Company, supra, 273 F. Supp. at pp. 594-595). As the Johnson court explained ( ibid. ): "If it has been necessary to open this Pandora's box [of state of mind] in the sphere of criminal responsibility and to apply our imprecise tools as best we can, most courts have, I think wisely, kept it shut ... by giving the `suicide, sane or insane clause' its plain, intended meaning. As the Supreme Court sensibly pointed out almost a century ago [in Bigelow ], it is precisely to avoid such imponderables that the insurer inserts the modifying language."

A construction of the clause as excluding all nonaccidental self-destruction regardless of the insured's state of mind is not only in accord with its plain meaning and manifest purpose, it is also supported by the great weight of authority interpreting "suicide, sane or insane" and similar phrases. (See, e.g., Bigelow v. Berkshire Life Ins. Co., supra, 93 U.S. 284, 286-288 [ 23 L.Ed. 918, 919-920 ]; Dennis v. Union Mut. Life Ins. Co., supra, 84 Cal. 570, 571-572 ; Nielsen v. Provident Life and Acc. Ins. Co., supra, 596 P.2d 95, 97-98 ; Atkinson v. Life Ins. Co. of Virginia, supra, 228 S.E.2d 117, 120-121; Ann Arbor Trust Co. v. North American Co. for L. & H.I. (E.D.Mich. 1974) 383 F. Supp. 310, 312-315 , affd. and remanded with instructions (6th Cir.1975) 527 F.2d 526 , cert. den. (1976) 425 U.S. 993 [ 48 L.Ed.2d 818 , 96 S.Ct. 2206 ] [applying Mich. law]; Johnson v. Metropolitan Life Ins. Co., supra, 273 F. Supp. 589, 592-594 ; Aetna Life Insurance Company v. McLaughlin, supra, 380 S.W.2d 101, 102-106 ; Gibson v. Reliance Life Ins. Co. (1934) 172 S.C. 94 [ 172 S.E. 772, 774-775 ]; State ex rel. Shoemaker v. Daues (1925) 312 Mo. 62 [ 278 S.W. 735, 736-737 ] [applying Ill. law]; Silliman v. International Life Ins. Co., supra, 174 S.W. 1131 , 1133; Campbell v. Order of Washington (1909) 53 Wash. 398 [ 102 P. 410, 413 ]; Moore v. Northwestern Mut. Life Ins. Co. (1906) 192 Mass. 468 [ 78 N.E. 488, 490 ]; Clemens v. Royal Neighbors of America (1905) 14 N.D. 116 [ 103 N.W. 402, 403-404 ]; Seitzinger v. Modern Woodmen of America (1903) 204 Ill. 58 [ 68 N.E. 478, 481-482 ]; Scherar v. Prudential Ins. Co. (1902) 63 Neb. 530 [ 88 N.W. 687, 688 ]; Cotter v. Royal Neighbors of America (1899) 76 Minn. 518 [ 79 N.W. 542, 543 ]; Haynie v. Knights Templars' and Masons' Life Indemnity Co. (1897) 139 Mo. 416 [ 41 S.W. 461, 464-465 ]; Spruill v. Northwestern Mut. Life Ins. Co. (1897) 120 N.C. 141 [ 27 S.E. 39, 40-41 ]; Tritschler v. Keystone Mut. Ben. Ass'n., supra, 36 A. 734, 734-735 ; Billings v. Accident Ins. Co. of North America (1892) 64 Vt. 78 [ 24 A. 656 , 657]; De Gogorza v. Knickerbocker Life Ins. Co. (1875) 65 N.Y. 232, 237-238 ; cf. Minn. Stat., § 61A.031 (1982) ["The sanity or insanity of a person shall not be a factor in determining whether a person committed suicide within the terms of an individual or group life insurance policy regulating the payment of benefits in the event of the insured's suicide ...."].)

It is true that a few cases do support a different reading of the clause viz., that nonaccidental self-destruction is not within the suicide exclusion if the insured is "so insane" as to be unable to understand the physical nature and consequences of his act. This is the so-called "Kentucky rule" named apparently after the jurisdiction in which a large percentage of the cases supporting it was decided. (See Annot., Insurance: Construction of "Sane or Insane" Provision of Suicide Exclusion (1966) 9 A.L.R.3d 1015 , 1025-1026 [hereinafter Annot.].) I can discover no reason, however, why we should follow the Kentucky rule. Nor, to conclude from its silence, can the majority.

First, the cases supporting the Kentucky rule are in the distinct minority and are mainly older. (See, e.g., Christensen v. New England Mut. Life Ins. Co. (1944) 197 Ga. 807 [ 30 S.E.2d 471 , 153 A.L.R. 794 ]; Muzenich v. Grand Carniolian Slovenian Catholic Union (1941) 154 Kan. 537 [ 119 P.2d 504 , 138 A.L.R. 818 ]; National Life Ins. Co. v. Watson (1922) 194 Ky. 355 [ 239 S.W. 35 , 35 A.L.R. 156 ].) Before we join such company, we are entitled to sound and persuasive reasons for doing so. The majority give us none.

Second, the minority rule rests on a construction of the term "suicide" that is not in accord with the popular understanding, and therefore frustrates the reasonable expectations of the insurer and especially the insured. "The majority view" which I would follow "adopts a plain meaning and popular definition of the term as covering any act of self-destruction. The minority" which the majority here propose to adopt "views the term in essentially a criminal law or technical concept holding that understanding and intent are deemed essential elements of a suicide." ( Nielsen v. Provident Life and Acc. Ins. Co., supra, 596 P.2d 95, 98 ; accord, Aetna Life Ins. Co. v. McLaughlin, supra, 380 S.W.2d 101, 102 .) Yet we are obliged to interpret the words of the suicide exclusion clause, like all contractual language, "in their ordinary and popular sense, rather than according to their strict legal meaning" (Civ. Code, § 1644). We are also required to construe the clause, like all policy provisions, in such a way as to "effectuat[e] the reasonable expectations of the ordinary applicant" ( Smith v. Westland Life Ins. Co. (1975) 15 Cal.3d 111, 122 [ 123 Cal. Rptr. 649 , 539 P.2d 433 ]). The minority rule violates these basic principles of contract and insurance law.

Third, and most important, the Kentucky or minority rule is antithetical to the purpose of the suicide exclusion clause generally and the phrase "suicide, sane or insane" specifically. That purpose is to make the insured's state of mind largely immaterial and thereby "effectively remove[] the necessity of grappling with the `shadowy and difficult to define' line between sanity and insanity...." ( Nielsen v. Provident Life and Acc. Ins. Co., supra, 596 P.2d 95, 97 , quoting Bigelow v. Berkshire Life Ins. Co., supra, 93 U.S. 284, 287 [ 23 L.Ed. 918, 919 ]; accord, Johnson v. Metropolitan Life Insurance Company, supra, 273 F. Supp. 589, 594-595 .) In stark contrast, the minority rule requires even greater mental gymnastics in every case: under that rule, the issue whether an act of nonaccidental self-destruction falls within the suicide exclusion clause turns on such subtle distinctions as "the degree of insanity affecting the destructive act, which requires an investigation of the actual state of the insured's mind at the time, with the result that if it is found that he had sufficient mentality to realize that the act performed would result in his death, and intended such a result, the death falls within the exclusionary clause although the motivation to kill himself may have arisen from his insane state, but that if he was so insane as not to realize that the act would be destructive to him, or to intend such a result, the act cannot be classified as `suicide' for purposes of applying the exclusionary clause." (Annot., supra, 9 A.L.R.3d 1015 , 1018, fn. 9.)

Thus, the minority rule does more than merely reintroduce the very difficulty that the phrase "suicide, sane or insane" is intended to remove it makes that difficulty more difficult still. If as is surely true "the line between sanity and insanity is often shadowy and difficult to define" ( Bigelow v. Berkshire Life Ins. Co., supra, 93 U.S. 284, 287 [ 23 L.Ed. 918, 919 ]), how much more so must be the line between the minority rule's "varying degrees or aspects of insanity." ( Aetna Life Insurance Company v. McLaughlin, supra, 380 S.W.2d 101, 106 ; see Johnson v. Metropolitan Life Insurance Company, supra, 273 F. Supp. 589, 594-595 .)

To summarize: the clause excludes all nonaccidental self-destruction regardless of the insured's state of mind; intent is therefore immaterial, and the trier of fact need determine only whether the act of self-destruction, objectively viewed, was accidental. In other words, "[i]f the act of self-destruction would be regarded as suicide in the case of a sane person, it would be so treated as to an insane insured...." ( Atkinson v. Life Ins. Co. of Virginia, supra, 228 S.E.2d 117, 120 .)

Here the trial court ruled that plaintiff bore the burden of proving that the insured did not commit suicide and that she could carry such a burden by showing the insured lacked the requisite mental capacity to form the intent to take his own life. In so ruling, as both the majority and I conclude, the court erred. Where the majority and I disagree is in the nature of the proceedings on remand. The majority would have the insurer show, to avoid liability, that the insured committed the act of self-destruction with suicidal intent but would apparently permit the beneficiary to recover in any event if she demonstrates that the insured was unable to understand the physical nature and consequences of his act. By contrast, I would have the insurer show simply that the insured's death which by stipulation of the parties resulted from an act of self-destruction was nonaccidental.

The difference between the majority's remand and mine is clearly substantial, and will make itself felt here and in the future. My view would entail a relatively simple and inexpensive inquiry, which is consonant with the plain meaning and purpose of the clause: was the insured's death self-inflicted and accidental? The majority would require more of the same empty inquiry into the insured's actual state of mind and his capacity to understand the physical nature and consequences of his act that has been pursued by the trial court twice, the Court of Appeal twice, and the majority now. Such a result would also and most importantly compel insurers to rewrite their standard exclusion clauses to except all death, accidental and nonaccidental, during the initial period of the policy, and thereby narrow coverage significantly.

Because the majority's opinion is fundamentally flawed in its analysis and seriously harmful in its consequences, I dissent.

[1] This court denied Allstate's petition for hearing on November 8, 1979, with Justices Clark, Richardson and Manuel voting to grant.

[2] At trial there was a dispute as to the statement made by Martin. The report of the San Diego Police Department and the coroner's report both indicated that Searle reported that Martin said, "I'm sorry, but I've had it."

[3] In relevant part, the trial court's instructions stated, "In this action, the plaintiff has the burden of establishing by a preponderance of the evidence all of the facts necessary to prove the following issue: that Martin Searle did not intend to take his own life.

"If you find by a preponderance of the evidence that Martin Searle did not have the mental capacity to intend to take his life, including awareness of the nature and consequences of his act of self-destruction at the time he shot himself, you will answer `No' to the first question on the special verdict form.

"If you find by a preponderance of the evidence that Martin Searle did not have the mental capacity to govern his own conduct at the time he shot himself, you will answer `No' to the second question on the special verdict form.

"If you find that at that time Martin Searle did have the capacity to intend to take his life, including awareness of the nature and consequences of his act at the time he shot himself, you will answer `Yes' to the first question on the special verdict form.

"If you find that at that time Martin Searle did have the mental capacity to govern his own conduct at the time he shot himself, your answer will be `Yes' to the second question on the special verdict form."

[4] Allstate argues, however, that Dennis v. Union Mut. Life Ins. Co. (1890) 84 Cal. 570 [ 24 P. 120 ] stands for the propositions that suicide and self-destruction are synonymous, that the clause limiting liability is unambiguous, and that any issue of state of mind is irrelevant. The synonymizing language in Dennis appears only in dicta and this court did not address the difference between suicide and self-destruction.

Plaintiff, in support of her position that insanity negates suicidal intent, cites Workers' Compensation Act cases construing Labor Code section 3600, which precludes recovery in cases where injury is "intentionally self-inflicted." These cases, of course, are not on point.

[5] The issue of intent was submitted to the jury only in the form of a question whether Martin had "the capacity to intend to take his life" when he shot himself. On retrial, the trier of fact should determine whether Martin in fact had the intent, not merely whether he had the capacity to have the intent.

[*] Assigned by the Chairperson of the Judicial Council.

[1] Similarly, Webster's Seventh New Collegiate Dictionary (1970) defines "suicide" as "the act or an instance of taking one's own life voluntarily and intentionally esp. by a person of years of discretion and of sound mind. ..." ( Id., at p. 879, italics added.)

Although dictionary definitions might not always provide accurate indicators of common meaning, they are certainly superior to century-old legal opinions, such as that relied upon by the majority. (See post, at pp. 447-448.)

[2] The majority implicitly acknowledge this contradiction. Under their approach, the beneficiary may negate a showing of intentional self-destruction by proving that the deceased was "unable to understand the physical nature and consequences of his act." (Maj. opn., ante, at p. 429.) In common parlance, this "inability" is nothing more or less than one particular degree of insanity. (See 43 Am.Jur.2d (2d ed. 1982) Insurance, § 537, pp. 607, 608, fn. 29.)

Unfortunately, the majority obfuscate the issue in the present case by attempting to maintain a distinction between "insanity" and the inability of an insured to form the requisite intent to commit suicide. (See maj. opn., ante, at pp. 436-437.) Of course, by some definitions of insanity, an "insane" person may commit suicide. (See ante, at p. 432.) However, the issue in the present case is not whether a person who is "insane" in the abstract can commit suicide. The issue is, rather, the particular degree of insanity or lack of capacity that renders a person incapable of forming the intent to commit suicide within the meaning of the exclusion. (Cf. 39 Cal.Jur.3d Insurance Contracts, § 249, p. 493 ["It is universally recognized that self-destruction while insane is not suicide within a provision against death by suicide; the only conflict in the authorities is with reference to the degree of insanity which will nullify the provision"].)

[3] It is generally recognized that, in actual practice, most consumers do not even bother to read form contracts. (See Rakoff, Contracts of Adhesion: An Essay in Reconstruction (1983) 96 Harv.L.Rev. 1174, 1179 [hereafter Rakoff], and sources cited.) Hence, unambiguous contractual provisions may provide little more protection than ambiguous ones. Accordingly, it has been suggested that insurance contracts should be interpreted not solely on the basis of their terms, but also in light of the "reasonable expectation of the average insured." (Tobriner & Grodin, The Individual and the Public Service Enterprise in the New Industrial State (1967) 55 Cal.L.Rev. 1247, 1273 [hereafter Tobriner & Grodin].) One commentator has gone further, suggesting that since form contracts no longer indicate any meaningful consent by the consumer, all terms of such contracts should be presumptively invalid. (See Rakoff, supra, 96 Harv.L.Rev. 1174.)

Because the currently well-established doctrine of construing ambiguities against the insurer suffices to resolve the present case, I find it unnecessary to speculate on the applicability of these other theories.

[4] The court also reasoned that the words of the clause had a "precise, definite, wellunderstood meaning," and that "[n]o one could be misled by them...." ( 93 U.S. at p. 287 [ 23 L.Ed. at p. 919 ].) However, as the above-quoted dictionary definitions make clear, the clause is self-contradictory under the modern definitions.

[5] For example, one of defendant's experts defined "insanity" as the level of incapacity at which a person could be committed to a hospital against his or her will. That usage of the term bears at most a tangential relationship to an insured's capacity to form the intent to commit suicide. The determination of whether a person may be involuntarily institutionalized centers not on his inability to understand the nature of an act such as self-destruction, but on such factors as his lack of ability "to provide for ... basic personal needs for food, clothing, or shelter." (Welf. & Inst. Code, §§ 5008, subd. (h)(1); 5350, 5352.6.)

[1] Folk belief is that "you've got to be crazy to kill yourself." Yet suicide has been committed by persons in all strata of life and for motives rational and irrational. Demosthenes took poison to avoid capture by Macedonia, Hannibal to evade capture by Rome. Cleopatra is said to have chosen death from the bite of an asp, and Virginia Woolf took her life because she believed she was going mad. Celebrity status did not deter Jack London, Ernest Hemingway, James Forrestal and Marilyn Monroe. George Eastman, inventor of the Kodak camera, left a note, "To my friends: my work is done. Why not?"

[2] On the third question, I agree with the majority that the burden of proving suicide rests squarely on the insurer: it must show that the suicide exclusion applies and thereby exempts it from liability. ( Dennis v. Union Mut. Life Ins. Co. (1890) 84 Cal. 570, 571-572 [ 24 P. 120 ], overruled on another point in Zuckerman v. Underwriters at Lloyd's (1954) 42 Cal.2d 460, 474 [ 267 P.2d 777 ]; accord, Rome v. Life and Cas. Ins. Co. of Tennessee (La. Ct. App. 1976) 336 So.2d 275, 277 ; Adcock v. Life Assur. Co. of Carolina (1976) 31 N.C. App. 97 [ 228 S.E.2d 654, 656 ]; Life and Cas. Ins. Co. of Tennessee v. Daniel (1968) 209 Va. 332 [ 163 S.E.2d 577, 580 ].)

[3] Even whole life insurance insures against an unforeseeable event: the date of the insured's death.

[4] A suicide exclusion clause of limited duration will not, of course, avoid the costs of a planned suicide committed after the exclusionary period has expired. The insurer, however, evidently considers such costs, if any, insignificant or at least worth bearing in order to offer, as a selling point, broad coverage after the exclusionary period. The insurer, in other words, agrees to cover the risk of suicide, "provided the suicide occurs sufficiently long after the taking out of the policy to satisfy the company that it was not contemplated when the insurance was applied for." ( Longenberger v. Prudential Ins. Co. of America, supra, 183 A. 422, 426-427 .)

[5] As I shall show, this is the very result of the majority's misinterpretation of the clause.

[6] For evidence of how high such derivative costs can be, we need look no further than the case before us. Although it has not been disputed that the insured died of a self-inflicted gunshot wound to the head, this action has been litigated in the trial court, then in the Court of Appeal, then in the trial court again, then in the Court of Appeal again, is now here, and will return for a third time to the trial court. More particularly, after the Court of Appeal incorrectly held in Searle I that the suicide exclusion clause applied only to "sane suicides," it was necessary for the parties to conduct a four-day trial to determine whether the insured possessed the requisite mental capacity to form the intent to commit suicide, whatever that meant.