controls.44 A lifetime history of aggression differentiates adolescent
suicide victims from matched controls, even after accounting for dif-
ferences in mental disorders between suicides and controls.45
Other factors too point to an association of experience with vio-
lence and suicidality. Prison inmates are at increased risk for suicide
compared to community dwellers, and inmates completing suicide
are more likely to be incarcerated for manslaughter or murder as
compared to other prisoners.46 In these instances, as violence expo-
sure (as indexed by incarceration or manslaughter/murder convic-
tions) increases, so does suicidality, perhaps because violence expo-
sure lowers barriers to injury in general, self-injury in particular.
Of course, if the perspective presented here has merit, then those
who have habituated to pain and provocation through such means as
serious drug abuse and prostitution should have demonstrably high
suicide rates. Heroin users are fourteen times more likely than peers
to die from suicide, and the prevalence of attempted suicide is also
orders of magnitude greater than that of community samples. Of
course, several other perspectives might predict this association—for
example, maybe it is just that heroin overdoses are misclassified as
suicides. However, heroin overdoses per se appear to play a relatively
small role in suicide among this group.47
Regarding prostitution, a qualitative analysis of the narratives of
twenty-nine street youth involved in prostitution revealed that 76
percent of them had made at least one suicide attempt.48 The authors
of the study concluded that the experience of trading sex was heavily
72 ● WHY PEOPLE DIE BY SUICIDE
implicated in the youths’ suicidality. In another study, homeless
youth involved in prostitution were compared to homeless youth not
involved in prostitution; those involved in prostitution had more sui-
cide attempts than others.49
Of course, several psychiatric syndromes are at least partly defined
by behaviors that would be viewed here as incurring pain and pro-
vocation and thus potentially engaging habituation and opponent
processes—to name two examples, borderline personality disorder
(which involves serious behavioral impulsivity, including self-injuri-
ous behavior), and anorexia nervosa (which involves self-starvation).
The relevance of these and other disorders will be explored in a later
chapter, but for now, I’ll note that borderline personality disorder
and anorexia nervosa are among the most lethal of all psychiatric
disorders, with the usual mechanism of death being suicide.50
Daredevils—those who are thrill-seekers—are often injured, and
may be more prone to self-injury. Menninger believed that this con-
nection is explained by a death wish on the part of daredevils.51 I
have a different explanation—daredevils habituate to injury, includ-
ing self-injury, and thereby acquire the ability to enact lethal self-in-
jury. As to why daredevils are daredevils in the first place, I think that
personality traits like impulsivity, to be addressed in a later chapter,
are a more convincing explanation than a death wish.
Thus far, I have focused on those who have engaged in or observed
injury, pain, or violence and who are in some way under-privileged
or victimized. Though these studies are generally consistent with the
view that those engaging in or observing provocative behaviors expe-
rience more suicidality, it might be more convincing still if research
indicated that engaging in or observing provocative behaviors con-
ferred higher risk to suicidality even in high-privilege groups. This
may be especially persuasive for the additional reason that high-priv-
ilege groups may enjoy more protection from suicide than others
(e.g., through greater access to social and material resources and to
The Ability to Enact Lethal Self-Injury Is Acquired ● 73
mental health care); any increased risk in a high-privilege group is
thus not explained by lack of these protective factors.
Physicians are of interest in this regard. Through their training
and practice, physicians frequently observe the consequences of pain,
violence, and injury, and they gain specialized knowledge about le-
thal agents, dosing, methods of death, and so forth—that is, they de-
velop considerable competence and capability regarding suicide. In
this connection, Menninger observed regarding suicidal behavior,
“We physicians, familiar from our daily experiences with these un-
lovely sights, often forget that for most persons these barriers im-
posed by taboos are quite high.”52 A review of published original
studies concluded that physicians in general have an elevated risk for
suicide compared either to the general population or to other profes-
sionals.53
This is true for female and male physicians alike, but it is possible
that the discrepancy between female physicians and other women re-
garding this issue exceeds the discrepancy between male physicians
and other men regarding this issue. This differential discrepancy may
be mirrored in suicide rates. Suicide rates are particularly pro-
nounced in female physicians—that is, as compared to other women
(either in the general population or in other professional groups), fe-
male physicians have quite elevated rates of suicide, on the order of 3
to 5 times higher than other women. Rates are 1.5 to 3 times higher
for male physicians as compared to other men.54 Because men in
general have more opportunities than women in general to experi-
ence pain and provocation (e.g., through contact sports), it may be
more difficult for male physicians to outpace other men regarding
experiencing pain and provocation (thus there is a smaller discrep-
ancy between suicide rates in male physicians versus other men). By
contrast, the average female physician easily outpaces the average
woman with regard to experiencing pain and provocation (thus, ac-
cording to the present view, there is a large discrepancy between sui-
74 ● WHY PEOPLE DIE BY SUICIDE
cide rates in female physicians versus other women). It is also pos-
sible that female physicians’ relatively high suicide rates involve
gender-specific pressures at work.
An alternative explanation to the association between suicidality
and provocative and difficult experiences is that people who undergo
such difficulties become demoralized and hopeless, and because of
this, more prone to suicidality. This possibility is important to con-
sider (and is actually consistent with other parts of my model on
feeling a burden on others and feeling that one does not belong, dis-
cussed later), but there are aspects of the findings summarized above
that do not square very well with this alternative explanation. For ex-
ample, Grassi and colleagues assessed suicidal ideation in injecting
drug users, 81 of whom were HIV positive, 62 of whom were positive
for hepatitis C and HIV negative, and 152 of whom were negative for
/>
both HIV and hepatitis C. Suicidality scores were elevated in the
sample as a whole, but there were no differences among the three
groups in suicidality.55 If demoralization were the key mechanism,
one might expect the infected groups to display more suicidality; by
contrast, if a key mechanism is the provocative experience of re-
peated self-injection of illicit drugs (which all participants in this
study had experienced), one might expect equal suicidality across the
noninfected and infected groups (this was in fact the finding). Also,
given issues of status and privilege, the findings on physicians may
not be explicable through demoralization, although issues related to
burnout and job stress may be involved.
A second alternative explanation is that impulsivity—the tendency
to act without thinking—underlies and explains any relation be-
tween painful or provocative experiences and suicidality; after all,
impulsive people are, on average, more likely than others to experi-
ence various provocations and more likely than others to engage in
suicidal behavior. Indeed, it will be argued in a later chapter that
an impulsive personality style is conducive to the acquisition of the
The Ability to Enact Lethal Self-Injury Is Acquired ● 75
ability to enact lethal self-injury mainly because of the tendency of
impulsive people to experience various provocations. To address the
possibility that impulsivity explains the relation between provoca-
tive experiences and suicidality, studies are needed that account for
impulsivity in examining the relation between provocations and sui-
cidal symptoms. Few studies have explicitly taken this approach; nev-
ertheless, some extant findings are relevant. For example, studies
have shown that previous experience with suicidality predicts future
suicidality, even accounting for various indicators of impulsivity.56
Indeed, the “kitchen sink” study mentioned earlier documented a re-
lation between past and future suicidality, even when borderline per-
sonality symptoms—closely related to impulsivity—were statistically
accounted for. There are also clear cases of people with impulsive
personality features whose suicidal behavior was carefully planned
over days or weeks—Kurt Cobain’s suicide was of this sort.57 In these
cases, a direct influence of impulsivity on suicidal behavior is hard to conceive; by contrast, the current proposal that impulsivity indirectly relates to suicidality via the accrual of the capacity for lethal self-injury is compatible with the phenomenon of planned suicides in im-
pulsive people. Finally, it is hard to imagine that physicians are more
impulsive than the general population, yet, as documented above,
they have somewhat higher suicide rates as compared to the general
population. A view centered on impulsivity does not constitute a sat-
isfying explanation for elevated rates of suicide among physicians.
The model proposed here would predict that those who have faced
repeated violence, pain, or injury would, on average, experience
higher suicide risk (without necessarily having been suicidal before),
because their painful and provocative experiences will have engaged,
at least to a degree, the same habituation and opponent processes en-
gaged by self-injury. Studies on topics ranging from tattooing, to vio-
lence, to self-injected drug abuse, to suicide rates among physicians
can all be interpreted as consistent with the model.
76 ● WHY PEOPLE DIE BY SUICIDE
Pain Tolerance
If prior suicide attempts habituate people to provocation and pain, it
might be expected that their pain tolerance exceeds that of others.
That habituation to pain is implicated in suicidality is illustrated
starkly by some of the anecdotal evidence, described earlier. But is
there empirical evidence on this?
Israel Orbach and colleagues have reported that suicidal people
can tolerate extreme temperatures applied to the skin better than
other patients.58 This is termed a “thermal pain threshold,” and sui-
cidal patients tend to have higher thermal pain thresholds and higher
general pain tolerance as compared to controls. Additionally, in re-
sponse to electric shock, suicidal participants showed higher toler-
ance for pain and appraised the pain as less intense than psychiatric
control groups. Similarly, in another study, suicidal patients endured
more pain as compared to accident victims who had similar levels of
injuries.59 Suicidal inpatients show the highest thresholds for another
index of pain threshold, tolerance of pressure applied to the skin, as
compared to nonsuicidal inpatients and controls.60 Even among pre-
school children, some of whom had suicidal ideas and behaviors, the
suicidal children show significantly less pain and crying after injury
than does a psychiatric control group.61
Suicidal inpatients are not as physiologically reactive to a movie
on suicide as compared to controls.62 This would be consistent with
the idea that suicidal patients have gotten used to suicidal stimuli
and thus do not react to them as much as do nonsuicidal patients.
Remarkably, a large proportion of people with borderline personality
disorder who self-injure report no pain on self-injury, even in re-
sponse to considerable physical injury (e.g., deep cuts), and their
self-reports of no pain are supported by psychophysiological mea-
sures.63 Here again, it appears that with repeated experience, people
get used to self-injury, even to the point that it is not painful.
My colleagues and I found a similar result among adolescent psy-
The Ability to Enact Lethal Self-Injury Is Acquired ● 77
chiatric inpatients, many of whom had significant histories of suicide
attempts.64 Many also regularly self-injured (usually by cutting on
their arms or legs), not in an attempt to die, but rather, in an attempt
to feel better when they were emotionally distressed. We asked these
youths about the amount of pain they felt on self-injury. Almost half
reported that they felt no pain, even when fairly serious damage re-
sulted—another demonstration that people appear to get used to
even medically damaging injury. Others reported that they did feel
pain on self-injury. Very interesting in the present context, those who
felt no pain on self-injury reported an average of almost four lifetime
suicide attempts (self-injury when intent was death), whereas those
who did feel pain on self-injury reported a lower number of lifetime
suicide attempts (around two, on average). My interpretation of
these data is that those with more lifetime suicide attempts have ha-
bituated more than others to the pain of self-injury, so much so that
self-injury no longer causes them pain, even though they are engag-
ing in self-injury that would be quite painful to most people.
Though not on suicidality, J. R. Seguin and colleagues showed that
boys with a history of physical aggression were less sensitive to pain
(as measured by a finger pressure device) as compared to less aggres-
sive boys.65 Studies of this sort rai
se the possibility that pain sensitiv-
ity is suppressed by past self-injury as well as by engaging in other
provocative behaviors (like aggression). According to the model de-
veloped here, decreased pain sensitivity—whether gained through
self-injury or other provocative experiences like aggression—may re-
move a barrier to serious suicidal behavior.
There are people who have neurological conditions that render
them unable to experience pain. This is a serious condition, often in-
volving repeated injury and even death due to the person’s lack of
awareness of serious injury. This is a rare condition so that data on
suicidal behavior among such patients are unavailable. Even if they
were available, their interpretation would be clouded by the fact that
78 ● WHY PEOPLE DIE BY SUICIDE
these patients often die at early ages as a result of their condition.
Nevertheless, the very existence of the syndrome and the problems it
causes illustrate the value of at least some pain sensitivity, and the
dangers that emerge as people lose pain sensitivity.
It should be noted that this literature on pain sensitivity and
suicidality is relatively small, and that there is a lack of longitudinal
studies showing that pain tolerance is related to later suicidality.
Overall, however, it appears that those who attempt suicide, relatively
speaking, become buffered from physical pain and some other pro-
vocative stimuli, consistent with the current view that the trajectory
toward serious suicidality is characterized by increased ability to en-
dure pain and provocation.
Implications for Accrued Lethality
The nature of suicidal symptoms may change as experience with pre-
vious suicidal behavior accrues. That is, serious suicidal symptoms
(as compared to less severe suicidal symptoms) may become more
and more prominent with repeated suicidal experience. This begs a
key question—what represents “severe” versus “less severe” suicidal
symptoms?
Like others before us,66 my colleagues and I showed that all sui-
cidal symptoms are not the same and can be categorized into two do-
mains, which, while of course correlated, are discernible, and which
we named “resolved plans and preparations” and “suicidal desire and
ideation.”67
Why People Die By Suicide Page 10