WHAT
HAVE WE LEARNED ABOUT SUICIDE? A HARSH CRITIQUE OF THEORY AND RESEARCH
David
Lester
Stockton
University
The purpose of this essay is to review
the field of what is known as suicidology in order to explore what we
have learned about why people die by suicide. This review is going to be harsh,
not in terms of criticizing what researchers and theorists have written, but
rather by omitting citations to those who have become well-known in the field,
past and present.
I reviewed the research and theory on
suicidal behavior from 1897 to 1997 in four books which I titled Why People
Kill Themselves. Those books did not, however, critique the papers cited. The
aim was to provide for those in the field a convenient reference to what had
been published in those years. I did choose the outstanding contributors in the
3rd and 4th editions of the book, but many of those
chosen there are not cited here. They did publish a large number of articles on
one issue in the field, thereby stimulating others, but their contributions do
not necessarily assist our understanding of suicidal behavior.
This review will also focus on why
people die by suicide and not people who attempt suicide but survive.
The
Societal Suicide Rate
The major questions about the societal
suicide rate is what determines the variation over time within a region (time-series
theories) and what determines the suicide rate from region to region (ecological
theories). In fact, the explanations are typically given for both of these
variations.
There have scorers of correlational
studies on both of these variations, but the results are typically uninteresting
theoretically. For example, suicide rates (over time and region) are positively
associated with divorce rates and unemployment rates. This is hardly surprising.
Have there been any correlates of societal suicide rates that surprise us? No.
What is of use for understanding suicide?
Insight #1: Henry
and Short
One of the theories that has been
neglected is that provided by Henry and Short (1954). Henry and Short proposed
that, when people in a society are miserable or unhappy, if they have an external
cause to blame for their misery, then they will be angry, and homicide rates
will increase. If, however, there is no external cause to blame for their
misery, then they will blame themselves, become depressed and more likely to
kill themselves.
This theory explains, for example, why
the oppressed in a society have higher homicide rates while the oppressors have
higher suicide rates. In line with this, African Americans in the United States
have higher homicide rates while whites have higher suicide rates.
Interestingly, Henry and Short also
extended their theory and applied it at the individual level and, in addition,
proposed child-rearing techniques of punishment that would lead to individual
differences in suicidal-homicidal tendencies. They proposed that love-oriented
punishment techniques would lead children to suppress their anger and become
depressed and suicidal in the long term, while physical punishment would allow
anger to be experienced.
Insight #2:
Uematsue and the Cohort Effect
Uematsue (1961) in an obscure journal
proposed that the number of potential suicides in a cohort is fixed. If that
cohort has a high suicide rate at an early age, it will have a low suicide rate
later in life, and vice versa. Lester (1984) tested this hypothesis and partially
confirmed it. With data available for cohorts only for a maximum of 35 years
(rather than lifetime), Lester confirmed Uematsue’s hypotheses for female
suicide rates in the United States but not male suicide rates.
This raises the possibility that
suicide is, at least in part, genetically determined. Such a hypothesis could,
of course, be tested using methodologically sound twin studies, that is,
comparing identical with non-identical twin pairs. However, the childhood and
subsequent experiences of identical and non-identical twins are not similar.
Insight #3: The
Natural Suicide Rate
Maris (1981) speculated
that the suicide rate could never be zero no matter how ideal the
social-economic conditions were. Coming from a economic perspective, Yang and
Lester (2004) speculated that, just as the unemployment rate can never be zero,
perhaps the suicide rate of a society can never be zero. They tested this by
devising regression equations for several countries using divorce and unemployment
rates to predict the suicide rate over regions of each country. Setting the
divorce and unemployment rates to zero still predicted positive and non-zero
suicide rates for each country. For a review of recent research on this issue
see Yang and Lester (2021).
Insight #4: Taylor
and Moksony and a Broad Social Variable
Both Taylor (1990) and Moksony (1990)
argued that social indicators such as divorce rates do not directly affect the
suicide rate. Rather, such social indicators are measures of broader social qualities
which affect the suicide rate. Evidence for this comes from a study by Lester
(1995) who found that the divorce
rate of the states of the United States was associated with the suicide
rate of divorced people and also the suicide rates of single, married, and widowed
people, indicating that the divorce rate was an indicator of general social malaise.
Lester (1994) correlated many variables
over the states and found that one set of variables were highly
inter-correlated and the factor score (using a factor analysis) was associated
with the suicide rate of the states. The set included the crime rate, the
divorce rate, inter-state migration, the percent divorced, And alcohol
consumptions positively loaded on the factor and the percent born in state and church
attendance loaded negatively on the factor. Again, this suggests the presence
of a broader social variable such as social malaise or social disintegration.
Insight #5: The
Suicide Rate as a Random Walk
The random walk was a concept proposed by
Pearson (1905) and refers to a mathematical description of a path that consists of a succession of random steps. It describes such diverse phenomena as the path traced by a molecule as it travels in a liquid or a gas and the price of a fluctuating stock or stock index (Malkiel, 1973). If a stochastic process follows a random
walk, then any disturbance to the process will persist over time. If the time series suicide rate was a random walk, then the changes from
day to day would be random and yet subject to shocks, that is, societal crises.
The randomness might explain, in part, why suicide prevention efforts have
failed in the last ten years.
Yang (1994)
explored whether the suicide rate in the USA from 1933 to 1987 followed a
random walk. (The
USA started producing mortality statistics for the whole USA only in 1933.) Using econometric techniques of analysis, she found that the time-series
suicide rate fits a random walk process overall, for six age groups and for
four sex-by-race groups.
Insight #6: Suggestion
One societal
happening that might affect deviations from a random walk is the media coverage
of a celebrity suicide. Phillips (1974) and Stack (1987) have documented that celebrity
suicides result in an increase in suicides in the society after media coverage
of the suicide.
Insight #7: Access
to Methods for Suicide
There is more to the suicidal act than
the outcome. There is also the staging of the act (Lester & Stack, 2015)
which involves choosing the method for suicide, the location, whether to write
a suicide note, etc. Access to methods for suicide plays a large role. For example,
in the UK, when home gas for heating and cooking was coal gas, coal gas was a
popular method for suicide. When the UK switched to natural gas, which is less
toxic and so more difficult to use for death, the suicide rate declined. In
less industrialized and more rural countries, pesticides are used extensively,
as is their use for suicide.
Discussion
There have been other interesting
theories of the suicide rate in societies or in subgroups of the society which
have received only partial support (that is, support in some studies and a
failure to support the theory in other studies).
For example, the social deviancy
theory of suicide argues that suicide rates will be higher in groups in a
society which are deviant in age, sexual orientation, ethnicity, etc. For
example, Lester (1987) found that the fewer non-whites in a state, the higher
their suicide rate.
Subcultural theories focus on
characteristics of the society or part of the society. For example, states with
a southern subculture have higher rates of gun ownership and higher rates of
suicide using firearms (Lester, 1986-1987).
The
Individual Suicide
Insight #8:
Psychiatric Disturbance
It is obvious that an understanding of
suicide necessitates taking into account psychiatric disturbance. There are,
however, grave problems with psychiatry as a science and, therefore, with
psychiatric diagnosis.
To introduce you
to my major objection, let us assume you have a headache and a fever. You go to
your family physician, and he tells you that you have a disease called headache-fever,
or HF for short. What would you do? You’d run as fast as you could out of his
or her office and look for a good doctor. Medical illnesses are based on
causes. What is causing your fever? What is causing your headache? Is it caused
by a virus or bacteria? If so, which ones? Lyme’s disease or swine flu? Is it
because of a brain tumor and, if so, is it malignant or benign?
Psychiatric
disorders or mental illnesses are not defined by causes. They are defined by
clusters of symptoms. Let us say you are depressed. Maybe it is because you do
not have enough serotonin in certain regions of the brain. Maybe you have suppressed
and repressed anger felt toward significant others in your life so that you are
no longer conscious of the anger (a Freudian, psychoanalytic view). Maybe it is
because you have learned from your life’s experiences that you cannot get out of
the traps in which you find yourself (learned helplessness). Maybe it is
because that are not enough rewards (positive reinforcers) in your life, either
because you are in unrewarding relationships and employment or because you lack
the skills to obtain rewards from others (a learning theory perspective). Maybe
it is simply the melancholia that is part of all of our lives (Wilson, 2008)?
Don’t be fooled by
new revisions of the Diagnostic and
Statistical Manual (DSM). he criteria for this disease, disorders, or
illness (call it what you will) still do not involve causes! Part of the motivation for revising the DSM is that psychiatrists
cannot agree on which “illness” patients have. Using an older version of the
DSM, Beck, et al. (1962) found that four psychiatrists, individually interviewing
the same psychiatric patients, agreed only 54% of the time for the specific
diagnosis and only 70% for the major category (schizophrenia, affective
disorder, anxiety disorder, personality disorder, etcetera). In another study
of the older version of the DSM, Sandifer, et al. (1968) had psychiatrists in
three cities view tape-recorded interviews of psychiatric patients. In North
Carolina, the patients were more often labeled as having neurotic disorders, In
Glasgow, Scotland, the same patients were more often labeled as having
personality disorders, and in London (England) the patients were more often
labeled as having bipolar affective disorder (manic-depressive disorder)!
There have been
three modern critiques of the current psychiatric system. Robert Whitaker’s Anatomy of an epidemic: Magic bullets,
psychiatric drugs, and the astonishing rise of mental illness in America,
Irving Kirsch’s The Emperor’s new drugs:
Exploding the antidepressant myth, and Daniel Carlat’s Unhinged: The trouble with psychiatry. These books were favorably
reviewed by Marcia Angell, a former editor of The New England Journal of Medicine, a prestigious scholarly medical
journal, in The New York Review of Books
(June 23 and July 14, 2012). Loren Mosher, a prominent psychiatrist, resigned
from the American Psychiatric Association back in 1998, accusing the
association of selling out to the pharmaceutical industry that markets
psychiatric medications.
Even allowing for these problems with
the psychiatric diagnostic system, I have never come across a methodologically
sound study of psychiatric diagnosis of suicides. A sound study must have a
control or comparison group, AND the diagnosis must be carried blind as
to the presence of suicide or not. In all published studies, the diagnoses have
been made by psychiatrists knowing that the person was a suicide. This knowledge
accounts for some psychiatrists deciding that ALL suicides are psychiatrically
disturbed.
Insight #9:
Typologies of Suicides
There have been many typologies of
suicides proposed, but no one typology seems to be ideal. Van Hoesel (1983) chose
typologies proposed by ten scholars and had judges classify 404 suicides from
the files of the medical examiners in Baltimore and Maryland into the
categories listed in the ten typologies. There were 69 subtypes in these ten
typologies. Correlations between the 69 subtypes revealed five clusters of
subtypes.
Escape (90
suicides)
Confusion (52
suicides)
Aggression (77
suicides)
Alienation
(23 suicides
Depression/Low
Self-Esteem (112 suicides)
It is almost certain that no single
theory of suicide can explain all suicides, and so a meaningful typology
of suicides must be devised. Van Hoesel’s appears to be the best currently
available. Next, it may be possible to find the causes for suicide that are specific
for each type.
Insight #10: The
Moods of Suicides
In his book Turning
Points De Leo (2010) presented moving accounts from nine people who
attempted to kill themselves, but who survived. Eight of the nine were happy to
have survived. These accounts illustrate several features well-known to
suicidologists, such as escape from mental and physical pain (Sergio and Maria),
anger (Sandro), hopelessness and a feeling of being a burden (Anna), and
suicide soon after discharge from a psychiatric hospital (Fabrizio). It is
noteworthy that anxiety is noted by some. Alessa, Lucia and Maria all talk of suffocating
and not being able to breathe anymore, and two of them (Lucia and Maria)
indicate that the anxiety was long-standing and not simply a result of the
decision to kill themselves. In fact, Maria tried to die to get away from the
anxiety.
Most commonly, depression and, in
particular, the cognitive component of depression now labelled as hopelessness
are proposed as present in suicides (Beck, et al. 1974). To this we might now
add anxiety. Menninger (1938) classified the motives for suicide as to kill, to
be killed and to die. The emotions accompanying these motives are anger, guilt
and depression. Lester (1997) suggested that shame (as distinct from guilt)
also plays a role in some suicides.
This suggests that a typology based on
the emotions experienced by suicides could be of value.
Insight #11: A Two-Self
Theory of Suicide
Lester (2022) proposed a two-self theory
of suicide which he formally presented as a series of postulates and corollaries.
Lester defined a subself is defined as a coherent
system of thoughts, desires and emotions, organized by a system principle. For
this theory of suicide, it is assumed that there are two subselves, a suicidal
subself and a non-suicidal subself.
Postulate 1: Not every individual has both a suicidal subself and a
nonsuicidal subself.
Postulate 2: At any point in time, one subself is in control of the
mind. It may be said to have executive power.
Corollary 2a: When one subself has executive power, the other subself is
said to be suspended.
Corollary 2b: A subself may appear in many situations, or only on rare
special occasions. One subself may be domineering while the others is
submissive.
Corollary 2c: A subself may have executive power for anywhere from
seconds to hours or even longer periods of time.
Corollary 2d: Selfhood is whichever subself has executive power at the
time.
Corollary 2e: The existence of two subselves accounts for the inconsistency
in the behavior of individuals.
Postulate 3: Individuals
can seek to create new subselves for the future.
Postulate 4: The
individual can try to integrate the subselves.
Postulate 5: The suicidal subself may be a regressive subself developed
early in life, formed by the introjection of the desires and thoughts of
powerful others (in particular, parental figures) and imitation of their
personality and behavioral styles.
Postulate 6: The suicidal subself may be formed as a result of early
experiences.
Postulate 7: The possibility of
attributing negatively valued aspects (thoughts, desires, emotions, or
behaviors) of oneself to the suicidal subself may enable the individual to
maintain high self-esteem since the negative aspects of the suicidal subself do
not color the nonsuicidal subself.
Postulate 8: The two subselves may
become enmeshed, and the psychotherapist must help the client create
sufficiently impermeable boundaries so that the nonsuicidal subself can
withstand pressure from the suicidal subself to take over executive power and
resist intrusions from the suicidal subself into the nonsuicidal subself when
the nonsuicidal subself has executive power.
Postulate 9: It is possible to create new subselves such as mediators
and recording secretaries.
Discussion
Many theories of individual suicide
are not theories that are applicable to all suicides, but merely provide one
type of suicide and, therefore, should be included in typologies of suicide. For
example, Joiner (2005) proposed that suicides have perceived burdensomeness,
thwarted belonging, and the acquired capacity to inflict self-harm. However, Lester
and Gunn (2022) found that this theory applied to only 3% of a sample of suicides
that they studied. Perceived burdensomeness alone was found in only 15% of the
sample.
Another example comes from
Transactional Analysis which proposes that suicidal individuals were exposed to
desires of their parents that they not exist (Stewart & Joines, 1987).
Perhaps the infant is handled with disgust or perhaps the mother wishes that
the baby had never been born, a wish that may be expressed directly or
consciously transmitted to the baby.
A learning (or social learning) theory
of suicide proposes that the individual learned suicide from others. A famous example
is that both Ernest Hemingway and his brother Leicester Hemingway died by
suicide, following in the footsteps of their father’s choice to die by suicide.
Methodology
Insight #12: The
Method of Substitute Subjects and Suicidal Intent
The method of substitute subjects
is the study of attempted suicides in order to learn about suicides. Suicides
are deceased and cannot interviewed (of course), while attempted suicides can
be interviewed and given psychological inventories. Also, of course, attempted
suicides are of interest in their own right.
But, as Lester,
et al. (1975)
noted, to learn about suicides from a study of attempted suicides, the
attempters must be classified in terms of suicidal intent (or the
lethality of their attempt). For example, a sample of attempted suicides can be
divided into three groups, minimal suicidal intent, moderate suicidal intent and
high suicidal intent. Then, if some variable increases (or decreases) in a
linear fashion over these three groups, then extrapolation can be made to those
who died by suicide. (Correlational studies using a suicidal intent score can
also be used.)
Insight #13: Comparison
Subjects
This Insight is different because no
research study has ever appeared on this idea. Palmer (1960) conducted a study
of murderers by comparing them with their same sex siblings. Palmer also interviewed
the parents of the men – murderer and brother. The use of same sex siblings,
controlling for birth order of course, is that the subjects are matched on many
variables, such as family dynamics (especially between the parents) and social
and economic factors. Palmer was interested primarily in the physical and
psychological frustrations that the men had experienced in the childhoods and
adolescence, and he found that the murderers had experienced significantly more
frustrations and had developed much less effective coping mechanisms. A study
of suicides and their same-sex siblings would be extremely valuable!
Discussion
As mentioned in the beginning of this review
of what we know about suicide, the harshness is not in the comments about those
cited, but in the omission of scholars whom others may think have contributed
greatly to suicidology. What is apparent, however, is how little we know about
why people die by suicide.
For example, given an adequate
typology of suicides, have we a theory of causation of any type? One might
argue that we can list some risk factors. But risk factors and not necessarily
causes. For example, a risk factor for coming down with influenza is mingling
socially with others during a flue epidemic without taking precautions such as
wearing a mask or having a flu shot. But those are not the causes of influenza –
a virus is the cause.
References
Angell, M. (2012a). The epidemic of
mental illness: Why? www.nybooks.com/articles/archives/2011/jun/23/epidemic-mental-illness-why/?pagination=false
Angell, M. (2012b). The illusions
of psychiatry. www.nybooks.com/articles/
archives/2011/jul/14/illusions-of-psychiatry/?pagination=false
Beck, A. T., Ward, C H., Mendelson,
M., Mock, J. E., & Erbaugh, J. K. (1962). Reliability of psychiatric
diagnoses. American Journal of Psychiatry,
119, 351-357.
Beck, A. T., Weissman, A., Lester,
D., & Trexler, L. (1974). The measurement of pessimism: the hopelessness
scale. Journal of Consulting & Clinical Psychology, 42, 861-865.
Carlat, D. (2010). Unhinged. New York: Free Press.
De
Leo, D. (2010). Turning points. Bowen Hills, Australia: Australian Academic Press.
Henry, A. F., & Short, J., F,
(1954). Suicide and homicide. Glencoe, IL: Free Press.
Joiner, T. E. (2005). Why people
die by suicide. Harvard University Press.
Kirsch, I. (2010). The Emperor’s new drugs. New York: Basic
Books.
Lester, D. (1972, 1983, 1992, 2000). Why people kill themselves. Springfield, IL
Charles Thomas.
Lester, D. (1984). Suicide risk by
birth cohort. Suicide & Life-Threatening Behavior, 14, 132-136.
Lester, D. (12986-1987). Southern
subculture, personal violence (suicide and homicide) and firearms. Omega,
17, 183-186.
Lester, D. (1987). Social deviancy and suicide. Journal of Social Psychology, 127,
339-340.
Lester D. (1994). Patterns of
suicide and homicide in America. Commack, NY: Nova Science.
Lester, D. (1995). Is divorce an
indicator of general or specific malaise? Journal of Divorce &
Remarriage, 23, 203-205.
Lester, D. (1997). The role of
shame in suicide. Suicide & Life-Threatening Behavior, 27, 352-361.
Lester, D., Beck, A. T. &
Trexler, L. (1975). Extrapolation from attempted suicides to completed
suicides. Journal of Abnormal Psychology, 84, 563-566.
Lester, D., & Gunn, J. F. (2022). Is perceived burdensomeness present in the
lives of famous suicides? Death Studies, 46, 1801-1806.
Lester,
D., & Stack, S. (2015). Suicide as a
dramatic performance. New Brunswick, NJ: Transaction.
Malkiel, B. G. (1973) A random walk down Wall Street. New York: Norton.
Maris, R. W. (1981). Pathways to
suicide. Baltimore: Johns Hopkins University Press.
Menninger, K. A. (1938). Man
against himself. New York: Harcourt, Brace.
Moksony, F. (1990). Ecological
analysis of suicide. In D. Lester (Ed.) Current concepts of suicide, pp.
121-138. Philadelphia, PA: Charles Press.
Palmer, S. (1960).
A study of murder. New York: Thomas Crowell.
Pearson, K. (1905) The problem of the random walk. Nature. 72, 294.
Phillips, D. P. (1974). The influence of suggestion
on suicide. American
Sociological Review, 39, 340-354.
Sandifer, M., Horden, A., Timbury,
G., & Green, L. (1968). Psychiatric diagnosis. British Journal of Psychiatry, 114, 1-9.
Stack, S. (1987). Celebrities and
suicide. American Sociological Review, 52, 401-412.
Stewart, I., & Joines, V. (1987). TA today. Chapel
Hill, NC: Lifespace.
Taylor, S. (1990). Suicide,
Durkheim, and sociology. In D. Lester (Ed.) Current concepts of suicide,
pp. 225-236. Philadelphia, PA: Charles Press.
Uematsue, M. A statistical approach
to the host factor of suicide in adolescence. Acta Medica et Biologica, 8,
279-286.
Van
Hoesel, F. M. T. (1983). An empirical typology of suicide. Master’s
thesis, American University. Ann Arbor, MI: Proquest Dissertations Publishing.
Whitaker, R. (2010). Anatomy of an epidemic. New York: Crown.
Wilson, E. G. (2008). Against happiness: In praise of melancholy.
New York: Farrar, Straus & Giroux.
Yang, B. (1994) A random-walk hypothesis for
the suicide rate and its implications for Durkheim’s theory of suicide. In D. Lester
(Ed.) Emile Durkheim: Le Suicide 100
years later, pp. 319-324. Philadelphia, PA: Charles Press.
Yang, B., & Lester, D. (2004).
Natural suicide rates in nations of the world. Crisis, 25, 187-188.
Yang, B., & Lester, D. (2021). Is
there a natural suicide rate? An update and review. Suicide Studies, 2(4),
5-12.
