When people say this, what they usually mean is that suicide is so foreign to their minds that, in order to do it, they’d have to be psychotic, demented, intoxicated, or delirious. Indeed, in the next section on how death can merge into themes like belonging in the minds of suicide terrorists and others. I will point out that this merging requires a kind of break in the mind. Most who have not experienced this break find it extremely difficult to conceive of suicide, and so understandably attribute all kinds of motives and qualities to the suicidal mind to make it more understandable. One way of doing this is to imagine “they’re simply crazy.” Yet that is definitely not the case. This break in the mind is very specific; it has to do with breaking from the universal fear and revulsion that people have about death, coming instead to embrace and invite it, in a way, lovingly. Is this the same thing as being psychotic, demented, intoxicated, or delirious?

In a word, no. Anecdotes abound in which a person was seen or spoken with at one point in time, seemed coherent, calm, in tune with reality, and sober; and then was dead by suicide minutes of hours later. The high-school president who gave a lucid and rousing talk to students, faculty, and others on the future of the school, both for the short- and long-term, is a good example. This kind of performance from him was typical; no one noticed anything unusual or erratic about his appearance or behaviour. On the contrary, he was composed and articulate, just as he usually was, and then approximately four hours later, he jumped to his death from the upper floors of a downtown hotel.

It is conceivable that the principal’s mental status deteriorated markedly in the hours after the speech, just before his death. This strikes me as possible, but extremely unlikely. For one thing, those who are prone to dramatic deteriorations in mental status tend to show those tendencies early in life and relatively often. This did not appear to be the case with the principal. For another thing, there are numerous documented cases in which people have been in communication with family or authorities in the second and minutes leading up to their deaths by suicide. In the vast majority of these cases, those about to die were not incoherent or psychotic.

Filmmakers of the documentary The Bridge remarked that one of the most surprising things about the entire experience was that they could not tell who was who. Passersby on the bridge who seemed obviously distraught, unusual, or erratic would concern the filmmakers, leading them to call authorities. When authorities arrived, virtually none of these individuals seemed at risk. By contrast, those whose deaths were filmed shocked teh filmmakers, because seconds before, they appeared to be just one of the bridge’s many day-to-day pedestrians. Were it the case that those about to die by suicide are always psychotic or otherwise “out of their minds,” the filmmakers would have been better at discerning those at risk.

Physician John Gray, editor of the American Journal of Insanity from 1854 to 1886, held this same view. In the journal, he stated that “in the large proportion of cases, if not the majority, [suicide] is committed by sane people.” Much turns here on the definition of “sane”; it seems clear that what Gray meant was that most who die by suicide are not delusional or otherwise psychotic.

People who die by suicide do, I believe, undergo a kind of mental break involving their views of death – they come to see death as a comfort to others and to themselves – but this mental break is not the same thing as psychosis, intoxication, dementia, or delirium. Am I arguing then that people who die by suicide do not have mental disorders? Emphatically no. The best evidence to data indicates that around 95 percent of those who die by suicide have a diagnosable mental disorder at the time of their death. Personally, I believe that is probably an underestimate, and that the figure is closer to 100 percent – I believe that, in the vast majority of cases, the 5 percent without clear mental disorders at the time of their death are experiencing subclinical variants of mental disorders.

The claim, then, is that virtually all suicide decedents were experiencing mental disorders at the time of their deaths, but that relatively few were psychotics, demented, intoxicated, or delirious at the time of their deaths (some are, it should be emphasised; a point to which I will return). This may appear to be a contradiction, but it is not.

Mental disorders are surprisingly common, but most people who have a mental disorder are neither psychotic, demented, intoxicated, nor delirious. I got an early glimpse of this as a graduate student, when we were testing out one of the first generations of computerised psychiatric interviewing. We have several hundred undergraduates undergo the computerised interview, and the computer used preset algorithms to diagnose the students. Overall, the computer gave diagnoses to approximately 45 percent of the students … students who were selected at random from large subject pools. Our knee-jerk reaction was “that’s too high, something is wrong without algorithms.” But that wasn’t the problem at all. The fact is that when all mental disorders are surveyed, including those related to substance abuse and dependence and specific phobias, the percentages add up. The students legitimately met criteria for conditions like alcohol abuse and the various anxiety disorders, yet virtually none of the students who met criteria for these disorders was psychotic, intoxicated, demented, or delirious. Subsequent surveys of the U.S. general population have returned similar findings.

At any given time, around 5 percent of the U.S. population is experiencing major depressive disorder. The disorder involves sadness, insomnia, loss of energy, and the like, and it causes serious distress and affects people’s lives negatively. But in the majority of cases, it does not involve psychosis, dementia, intoxication, or delirium.

Perhaps an even more interesting example is substance use disorders. In the most serious version of these conditions, substance dependence disorders, people are clearly addicted to a substance, they have developed tolerance for the substance, they have prominent withdrawal from the substance in its absence, and the condition negatively affects their lives in clear ways (e.g., they don’t work because they are drug-seeking instead). Even among the subset of people with substance dependence disorders, most do not spend the majority of their time intoxicated. Even when using the substance in question, many are not particularly incoherent, because they have been using long enough that their bodies have adapted to the substance. This is not to say, of course, that alcohol and drugs cannot make people incoherent.

Many people have witnessed or experienced firsthand the cognitively incapacitating effects of severe alcohol intake, and the results can of course seem “crazy” indeed. Those without these firsthand experiences have no doubt seen them depicted in various media. Is it this kind of “craziness” that leads someone to something as hard to understand as suicide, many may wonder?

To reframe the question, what is the blood alcohol level of the average person who has died by suicide? And what is the blood alcohol level of matched controls (i.e., the average person who has not died by suicide but otherwise resembles suicide decedents regarding factors like age, race, psychiatric condition, and so on)? In answer to the first question, a review that appeared in the June 2006 issue of the American Journal of Forensic Medicine & Pathology is informative. The review looked at over 200 deaths by intentional overdose that occurred in Kentucky over a ten-year span. The majority, 67 percent, had no alcohol in their systems. A study from Australia on suicide attempts mentioned earlier indicated that 29 percent of attempters had been drinking at the time of their attempt. It appears that the majority of people are not even drinking at the time of their attempt or death by suicide, much less intoxicated.

Tom Hunt’s book Cliffs of Despair relates an interesting anecdote about the involvement of alcohol in death by suicide. Hunt reviewed the coroner’s file of a man who had jumped to his death by suicide off the cliffs of Beachy Head in England. In his suicide note, the man stated that although he had recently celebrated seven years of sobriety, the coroner will undoubtedly discover alcohol in his system. “He wants to make clear, though, that it wasn’t the booze that drove him to suicide. “It was clinical depression, existential angst, cosmic alienation or whatever”” (p. 165). Indeed, this man had arranged for his funeral years earlier, aware that he night eventually die by suicide. This man’s blood alcohol level was above zero as he died by suicide, but to imagine that alcohol caused his death seems far-fetched. Indeed, the man’s death shows another potential role for alcohol in suicide – to brace people against its fearsome and daunting qualities.

Another principle illustrated by the man’s death is that “above zero” alcohol level by no means ensures intoxication. From the Kentucky study noted above, about a third of suicide decedents had some alcohol in their blood at the time of death. Of these, most had blood alcohol levels below .08, the legal limit for driving under the influence in the United States. The man who died at Beachy Head almost certainly fit this description, and my dad certainly did. My dad’s blood alcohol level in his autopsy report was .02. My guess is that he had had three or four drinks a few hours before his death, perhaps in an unsuccessful attempt to take the edge off of his sleepless depression. Or perhaps he had a drink, maybe two, in the hour or so before his death, as he was contemplating what he ultimately would do. In either scenario, the “raving intoxicated madman” scenario of suicide did not fit my dad, as it does not fit most who die by suicide. In fact, even for those whose blood alcohol is above the limit of .08, not all of them are raving or even particularly intoxicated. By law, they should not be driving, but most are not “out of their minds.” For many of those who drink regularly, their behaviour at a blood alcohol of .10 is not much different from their behaviour when their level is zero; for some of them, they can seem calmer and more lucid at a level of .10.

Still another explanation for the association between alcohol and suicidal behaviour is that they are related to each other mostly because they both relate to a third variable. Consider a 1991 study that appeared in the American Journal of Public Health. The study was on Navajo adolescents, and found that consumption of hard liquor was associated with past suicide attempts in a linear way. That is, those who never consumed hard liquor had relatively low rates of past suicide attempts; those who drank hard liquor at least monthly had higher rates; and those who drank hard liquor at least weekly, the highest rates of all. If the story stopped there, it would not be particularly interesting, not would it be especially relevant to the third variable issue mentioned above.

But the story continues in two interesting ways. First, the same pattern of results on alcohol intake and suicide attempts did not pertain when beer or wine were examined instead of hard liquor. Or, to put it a little more accurately if complexly, the same pattern only held among beer and wine drinkers who also drank hard liquor. There was no association between beer and wine drinking and past suicide attempts once hard liquor consumption was controlled for. Why would this association be specific to hard liquor versus beer and wine? Is hard liquor more facilitative than beer or wine, or is it that hard liquor drinking is an indicator – a signal – of a broader, underlying risk for suicidal behaviour?

I think it is the latter, and this gets to the second compelling finding from this same study. As already noted, a strong predictor of past suicidal behaviour was the participant’s own use of alcohol, specifically hard liquor. But there was another alcohol-related index that outperformed even the participant’s own report of hard-liquor drinking – namely, the participant’s mother’s use of hard liquor. The effect of the participant’s father’s hard-liquor drinking was detectable too, but was weaker than both the participant’s report of his own liquor drinking and the mother’s report of her liquor drinking.

This is a fascinating finding. Liquor drinking is predictive of past suicidality – interesting enough but not too surprising. But that your mother’s drinking is predictive, not of her past suicidality, but of yours, and that it is more predictive of your suicidality even than your own liquor drinking – now that is interesting and surprising. That your father’s drinking predicts your suicidality but less powerfully than your own or your mother’s drinking adds another layer to these results.

What explains this intriguing set of findings? It seems not to be well explained by the idea that alcohol intake is a facilitative factor in suicide in the moment. If it were, why would liquor drive the effect and not wine or beer? Perhaps one could argue that it is simply that liquor is more intoxicating in the moment … but then how to explain that one’s mother’s liquor drinking is more predictive than one’s own? Whether this effect is genetic or related to family conflict (extreme alienation from family and community was also a robust predictor), it does not clearly support a proximal facilitative role for alcohol use; it is more consistent with the view that alcohol use is a signal of a deeper substrate of chronic risk – a risk that is passed on from parents to children. And why would a mother’s drinking and not a father’s drinking be predictive? Probably because it represents a clearer signal. That is, if one’s father drinks excessively, that can be a problem, but relative to other men, a lot of whom have the problem, the underlying condition is probably not severe. But if one’s mother drinks excessively, that almost certainly will be a problem, because not a lot of women do that; for her to do so means that she has a severe underlying condition, and that severity is getting signaled to you either genetically or through family environment.

Overall, the role of alcohol use in suicidal behaviour seems to be much more complex than it is usually portrayed. The situation is similar to that involving impulsivity and suicidal behaviour. There is a documented empirical association, but such an association can mean many things. It can conceivably mean that impulsivity leads people to die by suicide on the spur of the moment, and though this idea seems to have caught on, it is almost surely wrong. Impulsive people can have rough lives – accidents, arrests, fights – precisely because they are impulsive. People who regularly go through such experiences toughen up. Should they develop ideas about suicide, and some do because of difficulties like accidents, arrests, and fights, then, more so than others, they have the wherewithal to enact those thoughts.

The logic is very similar regarding alcohol use – indeed, there is an empirical association between impulsivity and alcohol use. If one is prone to use alcohol excessively, difficulties like fights, arrests, and accidents are more common, as are other demoralising incidents like relationship loss and job failure. It is this web of demoralising events and development of a tough shell that links alcohol use to suicidal behaviour, rather than alcohol making people suicidal in the moment.

Having said that, it is the case that impulsive people can get very emotional and thus act rashly, including acting in suicidal ways. In the section on impulsivity and suicide, I argued that even in this scenario, impulsive people are not dying by suicide on the spur of the moment; rather, they are impulsively resorting to a plan that they have in reserve. The plan “on the shelf” has the more prominent causal role in death in these cases, though impulsivity played some role too.

Alcohol use can function similarly. I have already noted that some people, perhaps including my dad, use it as a way to brace themselves for the very fearsome thing that they are contemplating. Another role for alcohol in suicidal behaviour involves what my FSU colleague Al Lang and others have termed “alcohol myopia.” This psychological state involves an alcohol-induced “tunnel vision” that focuses on only relatively salient and immediate cues and stimuli. For most people who are intoxicated, their myopia doe not predispose them to anything having to do with suicidal behaviour because they don’t have any suicidal plans. This is true even for people who are prone to negative emotions when drinking. But for a subset of people who have pondered suicidal behaviour over months or years, alcohol intoxication can narrow their field of view, such that a crisis focuses them on suicide. In this scenario, alcohol intake is playing a role, but not a primary causal one. Rather, intoxication is a contextual factor, a setting condition. Acute alcohol intake can focus people on only one or a few plans of action; for people who are vulnerable to suicidal behaviour to start with, a plan of action can involve death by suicide.

Although there is little doubt that some suicide decedents are intoxicated at the time of their death, this obscures two important facts: 1) most are not; and 2) most of those who are occasionally or even regularly intoxicated to do not attempt, much less die, by suicide. This same general line of reasoning applies to other states in which people’s mental status is impaired. That is, some people who die by suicide are psychotic (out of touch with reality, having hallucinations, delusions, and cognitive confusion) at the time of their death, but most are not, and most psychotic episodes do not include suicide attempts. Those with other impaired mental statuses, including various forms of dementia (that is, severe cognitive decline that includes memory loss) and delirium (a state of serious mental confusion), behave similarly.

In fact, those few individuals who do cause their own deaths while psychotic, demented, or delirious represent a challenge to frameworks that attempt a nomenclature for suicidal thoughts and behaviours. One such framework was put forth by Silverman and colleagues in 2007 in the journal Suicide & Life-Threatening Behaviour. One of the categories in this nomenclature is “Undetermined Suicidal Behaviours,” which are self-inflicted potentially injurious behaviours where intent is unknown. This term is applied when the intent to die is unknowable due to the fact that an individual is psychotic, demented, or delusional. These behaviours are classified differently than clear suicide attempts and death by suicide. An example of an “undetermined suicidal behaviour” under the Silverman et al. guidelines would be an individual who has bipolar disorder and is delusional in the midst of a mixed episode – such episodes are characterised by a mixture of manic and depressive symptoms as well as anger and irritability. Friends report that he has been rambling about themes of religious self-sacrifice, Jesus coming down from “the air to the earth to save us all,” and so on. He is later found dead after falling to his death from the roof of his apartment building. In this example, intent is impossible to decipher. It is possible that the man intended suicide, but it is also possible that he intended something very different by jumping – specifically, to save others as a messiah. The latter idea is psychotic, but whether or not it is clearly suicidal is a bit of a conceptual conundrum.

As another intriguing example, I am aware of a legal case in which an older person has died by suicide. The person had accumulated wealth, and had changed his will months before his death to leave most of the money to an agency of his choosing, rather than to family members. A lawsuit resulted, claiming that the later suicide proved that the man was not competent to make decisions of that sort when he changed the will. One sympathises with the family – they have lost a loved one; to add to their distress, my sense is that such lawsuits are not particularly successful because it is simply not the case that those about to die by suicide suffer mental impairment in the same way as those declared incompetent for legal proceedings.

The mental state of those who are on the precipice of suicide is different from that of others to be sure, but it is the nature of that difference that is in question. Many think the difference can be explained with reference to various impairments in cognitive status such as intoxication or psychosis. Although it has intuitive appeal, it does not fit the facts. The majority of suicide decedents did not display impairments like psychosis, dementia, or delirium before their deaths. What they did display is difficult for most to understand. Just as many cannot fathom losing touch with reality, most cannot imagine a state of mind in which death seems an inviting comfort. Even though it may be hard for most to imagine, this, not intoxication or psychosis, is what suicide decedents experience.