Baechlers Theory of Suicide

THE BEHAVIORAL AND BRAIN SCIENCES (1980), 3,265-290

Printed in the United States ol America

Human suicide:
a biological perspective
Denys deCatanzaro
Department of Psychology, McMaster University, Hamilton, Ontario,
Canada L8S 4K1

Abstract: Human suicide presents a fundamental problem for the scientific analysis of behavior. This problem has been neither appreciated nor
confronted by research and theory. Almost all other behavior exhibited by humans and nonhumans can be viewed as supporting the behaving
organism's biological fitness and advancing the welfare of its genes. Yet suicide acts against these ends, and does so more directly and
unequivocally than any other form of rnaladaptivc behavior. Four heuristic models are presented here to account for suicide in an evolutionary
and sociobiological framework. The first model attributes suicide to the extraordinary development of learning and cultural evolution in the
human species. Learning may make human behavior so independent of biological constraints that it can occasionally assume a form entirely
contrary to the principles of biological evolution. The second model attributes suicide to a breakdown of adaptive mechanisms in extremely
stressful novel environments. The third model involves kin and group selection, arguing that in limited circumstances suicide may occur
because of beneficial effects it has on other, surviving individuals who share the suicidal individual's genes. The last model suggests that suicide
should be tolerated by evolution when it has no effect on the gene pool. This model holds particular promise in accounting for aspects of suicide
not attributable to culture. The evidence indicates that suicide is most common in individuals who are unlikely to reproduce and unable to
engage in productive activity; such individuals are least capable of promoting their genes. A complete explanation of suicide may derive only
from an analysis of its biological significance.
Keywords: suicide; self-destruction; sociobiology; evolution; altruism; life-threatening behavior
Human suicide, unlike almost all other behavior, acts directly reproduction will probably promote his genes. If an individ-
against the survival of the behaving organism. It consequently ual survives through reproductive age, producing and foster-
presents a special challenge for general theories of behavior ing children, any genes contributing to his health and repro-
and a particular one for current sociobiological theory. It is duction will survive into future generations. However, an
generally held that evolution selects for behavior that individual's behavior may contribute to the welfare of his
promotes the existence of the organism and its genes, yet genes despite the fact that he may not reproduce. This can
suicide, which acts directly against these ends, is not uncom- occur because other related individuals share some of his
mon. The purpose of the present discussion is to examine the genes, and his behavior may promote the survival and repro-
clinical and sociological literature on suicide from a biological duction of those individuals (see Hamilton 1964). According-
and evolutionary perspective. The anomaly that suicide pres- ly, the concept of "inclusive fitness" has been developed. The
ents will be outlined, and some heuristic models that might inclusive fitness of an organism encompasses not only his
explain this anomaly will be presented. The feasibility of individual biological fitness but also the fitness of his genes as
conducting research to test the validity of evolutionary they exist in other organisms. An individual may also behave
models will then be discussed, and some specific research for the benefit of unrelated individuals not sharing his genes;
strategies will be suggested. this has been explained through the concept of "reciprocal
altruism," wherein some tacit or explicit contract ensures
reciprocity (Trivers 1971).
Evolution and behavior An overview of human and other animal behavioral
patterns suggests that almost all such patterns readily fit this
Much attention has recently been given to the relation framework. Almost universally, organisms actively avoid
between behavior and evolution (see reviews by Barash 1977; circumstances that threaten their existence or health, unless
Dawkins 1976; Maynard Smith 1978; Parker 1978; Wilson they have not learned or are incapable of learning the
1975). It has been explicitly or implicitly held that, in general, negative contingencies of such circumstances. Conversely,
an organism's behavior is oriented toward promoting its organisms generally seek situations in which food, water,
genes. It is reasoned that behavioral predispositions conducive shelter, physical comfort, and the like are available to
to the welfare of the organism's genes will survive in natural improve their welfare; they also invest a considerable amount
selection, while less adaptive predispositions will be elimi- of time and energy in reproductive activity. It is not difficult
nated from the gene pool. This is not a teleological position, to explain such behavior. Through natural selection, orga-
attributing "purpose" to behavior, but rather suggests which nisms that behave in ways that are conducive to survival and
types of behavioral predispositions survive in natural selec- reproduction will pass on their genes, while the genes of
tion. Nor does it entail that an individual is "conscious" of the organisms that do not do so will be eliminated from the gene
effect of his behavior on his genes, but rather that he will pool. Thus any genes predisposing toward behavior that
probably inherit factors leading his behavior to advance his promotes survival and reproduction will have a selective
genes. advantage over other genes.
Any behavior supporting the individual's survival and In adopting these notions one does not assume that all
91980 Cambridge University Press 0140-525X/80/020265-26$04.00/0 265
https://doi.org/10.1017/S0140525X00004799 Published online by Cambridge University Press

deCatanzaro: Biology of suicide
behavior is under direct genetic control. Behavioral pheno- stereotyped, repetitive self-injurious behavior, which occurs
types may be strongly influenced by conditioning processes, in schizophrenic, retarded, and other psychiatric populations
which select for adaptive behavior within each individual and which may or may not bear some resemblance to suicidal
without requiring genetic change. The learning processes behavior (deCatanzaro 1978; 1979; deCatanzaro and Baldwin
themselves, rather than specific response topographies, may 1978). All of these behavioral patterns have as their major
have been selected by evolution; indeed, the uniformity of apparent consequence a self-produced decrease in biological
learning processes across individuals and species (cf. Hinde fitness.
and Stevenson-Hinde 1973; Nevin and Reynolds 1973) and In the present discussion, unless otherwise specified, the
indications of nonassociative physiological mediation (see term "suicide " refers only to completed suicide. Completed
Gallistel 1973; Mogenson and Phillips 1976; Stein 1978) suicide is perhaps the most unambiguous form of self-destruc-
suggest that such processes have an inherited, biological basis. tive behavior, in that it is difficult to conceive of any
Also, many motivational variables that interact with learning secondary gain that could accrue to an individual not surviv-
may be influenced by innate factors. This follows from ing his suicidal act.
evidence of nonassociative physiological mediation of aspects Suicide has been evident throughout recorded human
of feeding, drinking, sexual behavior, aggression, pain avoid- history (Farberow 1972). Records from early Western civili-
ance, and emotion (see Barchas, Akil, Elliott, Holman, and zation suggest that it occurred frequently among the Greeks,
Watson 1978; Carlson 1977; Conner 1972; Gorzalka and Romans, and early Christians, and perhaps to a lesser extent
Mogenson 1977; Schildkraut and Kety 1967; Stein 1978). It is among the Jews. In the Orient, it was reported in early
thus likely that behavioral development involves complex Brahman and Buddhist writings and condemmed with sever-
interactions between learning and innate variables; the innate ity in Muhammadan teachings. Suicide occurs in a wide
component should reflect countless generations of selection variety of human cultures today, with varying frequency,
for adaptiveness, and we should thus be able to account for including such diverse groups as North American Indians,
most behavior in terms of evolutionary contingencies that
affect it. Europeans, Chinese, Japanese, and East Indians (see Farbe-
row 1975).
Suicide currently accounts for a substantial portion of all
human deaths. Overall, it ranked as the ninth most frequent
The problem of suicide cause of death in the United States in 1975, occurring at a rate
of 12.7 per 100,000 population (Demographic Yearbook
Suicide is anomalous within this evolutionary perspective. In 1976-1977). Rates in many other countries are considerably
contrast to almost all other behavior, suicidal behavior acts to higher, and generally are greatest in the developed countries.
decrease rather than increase the biological fitness of the In 1975, the highest rates were in Hungary, with 38.4, and
behaving organism. Indeed, suicide is behavior operating Finland, with 25.1 per 100,000 population. Generally, men
directly against the behaving organism's survival. Since commit suicide more frequently than women, while the
behavior oriented toward propagating the organism's genes frequency of successful suicide increases as a function of age,
should generally promote the organism's existence, suicide being most common in the elderly (Diggory 1976; Linden
requires a special explanation in behavioral theory. and Breed 1976). However, suicide accounts for a larger
A cursory analysis of the problem might suggest that portion of all deaths in young adults, where in many countries
suicide should be very rare. Individuals committing suicide it is the third or fourth most frequent cause of death for that
remove their genes, at least those carried specifically by the age group.
individuals themselves, from the gene pool. Consequently, These statistics reflect only the deaths that are formally
any genes that they carry that permit or support suicidal registered as suicide. Many investigators believe that this
behavior should be selected against. Conversely, individuals represents a considerable underestimate (e.g. Stengel 1973).
possessing genes acting to reduce the likelihood of suicidal Many suicides may be registered as accidental deaths to spare
behavior should have a strong selective advantage over those relatives embarrassment, or because of uncertainty about the
lacking such genes. Given such selection pressures, suicide cause. Moreover, these statistics reflect only completed
should be at best an infrequent event. Yet suicide is a fairly suicides attributable to discrete acts. If death resulting from
common cause of death in humans, while a number of other, chronic subsuicidal behavioral patterns were included, the
"subsuicidal" behavioral patterns in humans, also acting rates for suicide would surely be much higher.
against the behaving individual's survival, may be even more
common.
Suicidal behavior occurs in diverse forms. While there have Models of suicide
been a number of attempts to develop nosological systems
(e.g. Durkheim 1951; Pokorny 1974; Shneidman 1968; 1969), On the basis of evolutionary theory and existing data on
there is no universally accepted framework for classifying suicide, some hypotheses that might explain the anomaly of
suicidal acts. Many investigators (e.g. Kreitman 1977; Stengel suicide can be suggested. These models outline potential
1973) feel tha,t the factors underlying completed suicide relationships among suicidal acts, pressures of natural selec-
differ from those underlying attempted but unsuccessful tion, and documented proximate determinants of suicide.
suicide, in that the latter involves methods with a low proba- They are developed here only for their heuristic value in
bility of success and may relate to reinforcement accruing to suggesting research strategies. The models explain suicide as
the surviving individual. Others (e.g. Menninger 1938; resulting from independence of learned behavior from natu-
Roberts 1975) have suggested that a broad range of life- ral selection, pathological conditions, actual selection favoring
threatening behavioral patterns, such as excessive risk taking, suicide, or evolutionary tolerance of this behavior.
alcoholism, and heavy consumption of other drugs, must be These may also be viewed not so much as models, but as
included in a definition of suicide. Menninger (1938) has. factors that could each contribute to the occurrence of
called these "chronic suicide," since the self-destructive suicide. There is no reason to assume that the factors are
behavior is dispersed over time. Shneidman (1969) has spoken mutually exclusive; rather, it is likely that some combination
of "sub-intentioned death," which involves the individual's is involved.
playing some partial, covert, or unconscious role in facilitat-
ing his own death; he has suggested that this may be charac- I: Independence from biological evolution. The first
teristic of a large number of human deaths. There is also model of suicide attributes this behavior to the human species'
266 THE BEHAVIORAL AND BRAIN SCIENCES (1980), 3

extraordinary development of the capacity to learn. Such technologies, while methods such as drowning, jumping from
learning ability may make behavior sufficiently flexible and high places, and exposure to the elements must have always
independent of biological evolution to allow the occasional been available to man.
maladaptive behavior, such as suicide, to arise. Furthermore, complex and unusual cognitive events
In many respects, human suicide may be unique in the frequently underlie suicide (see Levenson 1974; Neuringer
animal kingdom. There are only limited examples of suicidal 1976; Shneidman 1957; Tripodes 1976); these events
behavior in other species, and these may be qualitatively presumably result largely from an individual's learning
different from human suicide. Among insects, there are many experiences, and may be especially related to his learning
behavioral patterns in which the behaving insect dies in about death. For example, some individuals may believe that
protecting the social group to which it belongs. For example, following suicide they will enter an afterlife that is better
a honeybee worker will usually die when it stings. It is than their present life or that they will rejoin a lost relative.
believed that such self-sacrificing behavior occurs because it Other individuals may see their successful suicide as
improves the fitness of reproducing individuals sharing the producing desired effects on their environment, but not
genes of the suicidal individual (Wilson 1971). Among perceive clearly the relation of this act to their personal
nonhuman vertebrates, however, clear examples of suicidal fitness. Such suicides may be motivated by the desire to have
behavior are lacking. The reputed mass-suicidal behavior of an impact on the environment, such as improving the welfare
the Norwegian lemming may actually be a population of others, taking revenge on others, or increasing the respect
dispersion phenomenon in which death is incidental to the or attention received from others. Suicide may also be seen as
presence of major geographic barriers (see Clough 1965). a way of removing oneself from pain or other aversive events.
Some mammals, particularly primates, will, under limited Normal, non-life-threatening behavior directed toward these
circumstances, engage in chronic self-injurious behavior, but ends might increase the fitness of the behaving individual;
it is not established that this is suicidal (see deCatanzaro, but individuals committing suicide to achieve such ends do
1978; 1979). Another exception may have occurred in a case not, of course, survive to experience increased fitness. It is
recently reported by Jane Goodall (1979), in which one conceivable that, because of such aberrant cognitions or
eight-year-old chimpanzee became lethargic, ceased eating, logical errors, suicidal individuals falsely perceive that their
and died following the death of his mother, although this is behavior increases fitness when it actually has the opposite
subject to other interpretations. There may also be cases in effect.
many species in which organisms sacrifice their lives for Another indication of a mediation or modulation of suicide
offspring (see Hamilton 1964; 1970; 1971; Wilson 1975), by learning is the involvement of modeling, or learning
although human suicide also occurs in many other situations. It through the observation of others' behavior (Bandura 1971).
remains possible that suicidal acts occur in other vertebrates Suicides are often clustered in location and time. They may
but are undocumented because of a lack of thorough increase with reports of suicide in the news media or be
ethological and laboratory research. Current evidence, brought on by suicide among friends (Hankoff 1961; Phillips
however, suggests that behavior oriented directly toward the 1974). The involvement of one type of modeling process,
behaving individual's death is almost uniquely human. contagion, is illustrated by the recent mass suicide in
Human behavior in general can be distinguished from that Jonestown, Guyana. The simultaneous suicide of several
of other animals in a number of relevant respects, notably hundred individuals must indicate a major contribution of
communicative ability, complexity of the social structure, and learning and suggestion to the occurrence of suicide.
flexibility of the behavioral repertoire (Washburn 1978; However, this does not necessarily indicate that suicide is
Wilson 1975). These differences have made possible the exclusively determined by learning. Modeling may modify
development of cultural evolution, which permits changes in the probability of suicide and determine its temporal and
behavior, social patterns, and technology without the direct physical location. It may not, however, by itself provide
influence of biological evolution (see Alexander 1979; sufficient motivation for suicide. In the Jonestown case, for
Durham 1979; Irons 1979). Human behavior probably relies example, many of those attracted to the cult may have been
on learning to an extent not paralleled in other species. inclined toward suicide before joining. Stress and desperation
Learning allows changes in behavioral phenotypes within may thus also have been major determinants of their suicides;
each individual's lifetime, and thus engenders fine these factors will be discussed in more detail below.
adjustment of behavior to the environment. Many behavioral On a larger scale, the prescribed patterns of behavior in a
patterns are thus neither directly controlled by genetics nor culture may determine when, where, and in whom suicide
subject to natural selection. Consequently, occasional will occur. Previous suicides in specific circumstances may
behavior that might not be adaptive in a biological sense, and act as models, teaching members of a culture that suicide is
that could not arise through biological evolution, might occur appropriate if similar circumstances are encountered.
because of extraordinary learning experiences. Although suicide occurs in some form in most cultures, there
Accordingly, suicide may be an entirely learned behavior, are differences in the situational determinants and forms
and its especially high incidence in humans may be across cultures. For example, the custom of self-immolation of
attributable to an extraordinary development of learning a woman on the funeral pyre of her husband has been
capacity in this species. Skinner (1969) argued briefly that particularly common in India (Venkoba Rao 1975). In Israel,
suicide must result from learning. He reasoned that any suicide is almost equally frequent in females and males,
genetic mutation favoring suicide would quickly eliminate contrary to trends elsewhere (Headley 1975). In general,
itself, and that therefore suicide must result from ontogenetic suicide is relatively infrequent in predominantly Catholic
rather than phylogenetic contingencies. Much of the available countries, which may reflect either that church's strictures
evidence concerning suicide can be construed as support for a against suicide or a tendency to certify suicide under other
model based on learning and cultural evolution. modes in those countries (Farberow 1972). There are
At the very least, many of the methods employed to numerous other differences among cultures and subcultures
commit suicide must be learned. Most common methods with respect to suicide (see deVos 1968; Farberow 1972).
today, such as those involving guns, drugs, and the Although differences in the recording of suicide statistics
automobile (see McCulloch and Philip 1972; Stengel 1973) among cultures cannot readily be factored out, cross-cultural
depend on human technological innovations. These methods variation does strongly suggest that learning and culture are
and their effects would have to be learned. However, other among the major determinants of suicide.
methods, such as stabbing, burning, or hanging involve older Some evidence also suggests a role of learning in
THE BEHAVIORAL AND BRAIN SCIENCES (1980), 3 267

subsuicidal forms of behavior. In some cases, completed who appear to be "gifted," some stressful experience is
suicide may be the unintentional result of such subsuicidal usually found in these cases (Shneidman 1976b).
behavior, thus making the suicide a result of learning Also in agreement with this model is some evidence (see
processes. This might occur where attempted suicide, Lester 1972; Stengel 1973; Temoche, Pugh and MacMahon
attributable to reinforcement expected after survival of the 1964) that suicide is correlated with other psychological
act (see Kreitman 1977; Stengel 1973), inadvertently disturbances. Its frequency in psychiatric populations is
produces suicide. Also, more rarely, self-injurious behavior, generally higher than in the population at large, and it
which may be partly attributable to reinforcement appears to be especially associated with affective disorders.
contingencies (see Bachman 1972; Carr 1977; deCatanzaro, However, it is questionable whether suicide is universally
1978; 1979), can sometimes lead to death. related to some other conventional category of psychological
The first model, then, reconciles suicide with functional disorder. Stengel (1973) estimates that on the average only
analyses of behavior by maintaining that it results from one-third of the people who commit suicide have been
learning and culture. Since humans have a highly developed suffering from any neurosis, psychosis, or severe personality
capacity to learn, their behavior may be sufficiently disorder; the majority of people who commit suicide have not
independent from the constraints of natural selection that it been under any psychiatric treatment. It may thus be
can occasionally act contrary to this selection. Selective difficult to form any generalization about the relation
pressures favoring a high degree of behavioral flexibility may between suicidal acts and mental disorder without some
outweigh negative pressures accompanying occasional expansion of the latter concept.
incidental maladaptive behavior. Indeed, there are strong Another notion consistent with the present model is that
indications that learning is involved in the etiology of suicide. current high rates of suicide reflect the fact that modern
It can determine the method employed and the time and society makes excessive demands on many individuals. Since
place of suicide. It is less clear that the underlying these demands may be historically novel, many individuals
motivational events are entirely due to learning, although may not be equipped to cope and may consequently display
learning probably modulates such motivation. maladaptive behavior such as suicide. There is some evidence
of an increase in suicide in recent times (see Diggory 1976;
Shneidman 1976a), although the rise may not be dramatic
II: Suicide as pathology. The second model of suicide and may also reflect improvements in the accuracy of
would account for this behavior in terms of a poor fit between demographic data.
the organism's inherited capacities and the environment. In The greatest difficulty with the pathology model is that it
general, the genes that are passed on in evolution may have does not fully account for the transition between stress-
survived because they facilitate adaptation, but in unusual induced disorganization of behavior and the occurrence of
environments they may fail to do so. Furthermore, genes behavior oriented against survival. One possibility is that
frequently mutate or recombine, often producing cognition might become sufficiently disorganized under
maladaptive phenotypes. These processes may engender severe stress to enable learning variables to produce suicide.
pathological conditions that facilitate the occurrence of Another possibility is that mechanisms mediating
suicide. physiological pain may not adequately inhibit self-injury in
An individual may inherit genes that, in interacting with a times of severe stress; evidence indicates that pain perception
normal range of ecological conditions, would make him is markedly reduced during stress by endogenous chemicals
well-adapted. These genes would thus usually be favored in known as endorphins, which act on receptors in the brain (see
natural selection. If the individual is exposed to environments review by Barchas et al. 1978). In such ways stress may
outside of those normally experienced by members of the facilitate suicidal behavior; the actual occurrence of suicide,
species, his genotype may fail to ensure that his behavior is however, may require the interaction of these factors with
adaptive. The possession of a normally adaptive genotype other variables.
does not prevent an organism from being damaged by There are some other, limited conditions under which
physical injury, disease, or congenital malformation. Nor does genes may predispose an organism to behave maladaptively;
it prevent psychological malfunctioning because of an these may for convenience be included in this model.
impoverished early environment, severe social isolation, or Evolution involves genetic mutation and recombination as
extremely stressful experience. Such conditions may be so well as natural selection. Such processes produce
extreme and unusual that the organism does not have an heterogeneity and allow genetic change, but may also
inherited capacity allowing him to cope. Indeed, this is true produce maladapted organisms. The involvement of
by definition of any extreme stress (Selye 1973). Under mutation in suicide would seem quite unlikely, given the
extreme stress of any sort, then, disorganized and frequency of suicide, although it might be involved in
maladaptive behavior may develop despite the fact that the exceptional cases or associated with some forms of self-
individual's genes would predispose toward adaptive injurious behavior (Lesch and Nyhan 1964). Also, because the
behavior under more normal circumstances. Such adaptive value of any gene depends on its genetic milieu,
disorganization of behavior may facilitate suicide. maladaptive gene combinations might occur despite the
Concordant with this model is evidence that suicide bears a adaptiveness of component genes in other circumstances
fairly consistent relation to stressful experience. Prolonged (Caspari 1967; Herskowitz 1973). This could conceivably
stressful experience is almost ubiquitous in the backgrounds influence the occurrence of suicide in that, for example,
of suicidal individuals, while some exceptionally stressful polygenic control of affective conditions (see Gershon,
Bunney, Leckman, Van Eerdewegh, and DeBauche 1976)
experience is frequently an immediate antecedent of the act
might occasionally produce individuals prone to severe
of suicide (see Ferrence, Jarvis, Johnson, and Whitehead depressive disorders, which in turn may be conducive to
1975; Kiev 1977; McCulloch and Philip 1972; Stengel 1973). suicide.
Divorce, loss of a loved one, unemployment, poor health,
terminal illness, financial problems, alcoholism, and Although this model may not account for all aspects of
disordered personal relationships frequently constitute such suicide, it does appear to explain very well the chronic
stressful experience. Indeed, suicide is probably very rare in self-injurious behavior that occurs in primates raised in social
the absence of some sort of stress or severe difficulty in isolation and in autistic, schizophrenic, and retarded humans.
coping. While there are reports of suicide among individuals As I have discussed elsewhere (deCatanzaro 1979), such

self-injurious behavior is almost universally associated with such conditions.
physiological, genetic, and developmental abnormalities that Other evolutionary mechanisms can be invoked to explain
may directly produce the behavior. suicide. Wynne-Edwards (1962; 1971) has argued that
The present model, then, proposes that suicide may result behavioral traits may be affected by selection among groups
from a disorganization of behavior accompanying the or subpopulations within a species, because reproductively
occurrence of normally adaptive genes in unusual isolated subpopulations compete and have different gene
environments or the occurrence of novel genes or gene frequencies. Stanley (1975) has argued that selection above
combinations in more normal environments. the species level may induce more genotypic change than
does selection acting on individuals. Through selection at
Ill: Suicide as altruism. In the third model, I would like to these levels, genetic traits that, when expressed, decrease an
consider whether some of the selection processes favoring individual's chance of survival, might increase in frequency.
"altruistic" behavior, discussed by the sociobiologists, might This may occur in that some competitive advantage accrues
be involved in suicide. This model suggests that self- to that individual's population, which shares these traits,
destructive behavior may occur because it has beneficial despite or because of his loss. For example, in a human tribal
effects for individuals other than the suicidal individual. This grouping, the loss of members experiencing certain extreme
would imply that an individual's inherited motivational stresses and thereby taxing common resources might improve
constitution is such that he may occasionally sacrifice his own the general fitness of the tribe. Tribes in which burdensome
fitness for others, which might suggest that genetic factors members were lost through suicide might fare better than
could indirectly favor suicide. those in which suicide did not occur.
Prosuicidal genetic factors that were expressed in all Some notions of suicide might be subsumed by a group-
individuals possessing them would strongly select against selection hypothesis of human suicide. Durkheim (1951), in
themselves. However, such factors might survive and even his pioneering work on suicide, states that one form of this
prosper if expressed in conditions where they confer a benefit behavior could be classified as altruistic. Altruistic suicide
on surviving individuals sharing the genes of the suicidal occurred in individuals who felt very much part of a group
individual. If evolution selects only for behavioral dispositions and sacrificed themselves for the good of that group or of
favoring the individual, suicide may not have genetic society in general. Freud (1924), on the basis of clinical
underpinnings. But in considering the fact that an observations, postulated the existence of a "death instinct" in
individual's genes are shared by others, a genetic basis for humans which only surfaced under limited circumstances.
suicide becomes conceivable. The model does not require that Furthermore, there are indications that feelings of guilt and
such prosuicidal genes directly control the topography of other negative self-attitudes may be correlated with suicide
suicide, but rather that the inherited motivational constitution (Kaplan and Pokorny 1976). Negative self-concepts may be
of individuals be such that they could behave for the benefit created by the behavior of other members of society toward
of others at their own expense. these individuals; these self-concepts may help to elicit
Hamilton (1964; 1970; 1971) proposed that behavior suicide.
sacrificing the fitness of the behaving organism may develop Farberow (1972) has described "institutional" or "social"
through its benefits for related individuals. This process, forms of suicide, involving self-destruction that a society
which has been termed "kin selection" (Maynard Smith demands of a member. Institutional suicide has occurred in
1964), occurs when the loss of one individual supports the many forms throughout recorded history. Farberow discusses
survival of close relatives. Improved fitness of relatives is an example from a primitive tribe, in which an individual
sufficiently great that the loss of one individual's genes is accused of transgressing the tribe's taboos would climb to the
outweighed by enhanced fitness of identical genes in top of a palm tree and plunge headfirst to his death. Another
relatives. Kin selection might be adapted to explain suicide by example is found in other cultures, where in times of food
attributing it to prosuicidal genetic factors expressed when shortage, the old and sick were expected to sacrifice
the individual is or perceives that he is a burden on relatives. themselves for the good of the group. Also, throughout
Relatives would share genes directly or indirectly favoring history, in times of war men have jeopardized and sacrificed
suicide, but because their circumstances were better, these their personal existence for the benefit of the social group to
genes would not be expressed in them. which they belong. These forms of suicide are compatible
Kin selection probably could not explain the majority of with a group-selection hypothesis, but they might equally be
cases of suicide. Frequently, the suicide of an individual is the attributable to cultural evolution and may thus not involve an
source of considerable grief to surviving kin (McCulloch and elicitation of special suicidal tendencies.
Philip 1973), which is difficult to reconcile with the notion There is considerable controversy about the importance of
that this death is for their benefit. Furthermore, suicide often group selection in evolution (cf. Dawkins 1976; Wiens 1971;
occurs when an individual has little contact with his family Wilson 1975; Wynne-Edwards 1962; 1971). Currently, many
(Stengel 1973); in fact, protracted isolation from family and people appear to support the view that the role of group
other members of society is among the best predictors of selection has been relatively minor. For example, Dawkins
suicide (see Batchelor and Napier 1953; Breed 1972; Ganzler (1976) argues very strongly that the dominant force in
1967; Leese 1969; McCulloch and Philip 1972; Sainsbury evolution is individual selection, and that most of the animal
1955; Schneider 1953; Stengel 1973; Worden 1976). behavioral patterns attributed by Wynne-Edwards to group
Nevertheless, given that individuals may be motivated to selection can be explained in other ways. He has also argued
act for the benefit of kin, and given the variance in individual that selection due to the differential reproductive success of
motives underlying suicide, kin selection may play a role in a individuals proceeds much more rapidly than selection due to
minority of cases. This is suggested by references in some the differential success of whole breeding populations.
suicide notes to benefits accruing to surviving kin (see, for Therefore, traits causing individuals to increase their genetic
example, Shneidman 1976c, p. 270). Also, there are some representation in future generations will win out over traits
indications that certain family patterns, such as broken favoring the population or species. Any individual within a
homes, economic difficulties, and the consistent receipt of group who behaves in a self-serving manner when altruistic
blame for family problems, commonly occur in the behavior is called for will be likely to pass on genes supporting
backgrounds of suicidal individuals (Jacobs 1971; Richman self-serving behavior; through such processes individual
1971); it is conceivable that kin selection might operate in selection may be more powerful than group selection.

It may nonetheless be premature to dismiss entirely the role circumstances and expects little improvement in the future
of group benefit in human behavior. Most theorizing about (Bogard 1971). Farber (1968) has argued that suicide occurs
group selection has dealt with nonhuman animals. During when people envision not only the present but also the future
most of human history there appears to have been an as negative. He asserts that suicide is unlikely when people
extraordinary degree of intraspecific aggression organized at anticipate acceptable life conditions and have a sense of
a group level. While complete annihilation of one group by competence, but likely when acceptable life conditions are
another may have been rare, some differential reproductive seriously and permanently threatened. Indeed,
success would probably accrue from success or failure in such "hopelessness" is an excellent predictor of true suicidal intent,
competition. Furthermore, the human situation may be being much better than "depression, " which may be merely a
complicated by the coevolution of culture and biology (see transitory state (Beck, Kovacs, and Weissman 1975).
Durham 1979). Cultural factors have interacted with Human suicide occurs, then, in individuals experiencing
biological evolution in a complex manner that makes an extreme difficulties in coping and who expect little improve-
analysis of biological changes exceedingly difficult. ment in the future. Given our assumption that coping behav-
Also, much of the current theory of selfishness and altruism ior promotes the individual's genes, difficulty in coping
has been based upon concepts of gene sharing among would seem to entail difficulty in advancing the welfare of
individuals. The coefficients of relationship most commonly one's genes. Conversely, suicide is relatively rare in individu-
used (Hamilton 1964; Wright 1922) give a brother or parent als who are well adapted; well adapted individuals may, by
one-half of an individual's genes, a cousin one-eighth, and so definition, be those most capable of reproducing or otherwise
on. According to this model, one begins to approach a zero promoting their genes.
relationship rapidly in considering individuals progressively The difficulties in coping associated with suicide may be of
less related in a familial sense. Yet this takes into account only a diverse nature, but are generally severe enough to impede
gene sharing by proximate descent and not gene sharing by behavior that effectively promotes the individual's genes. For
kind. Also, it considers only a few generations rather than the example, suicide is frequently related to economic hardship,
totality of the species' history, which might produce much commonly occurring in the unemployed or those who have
more complex interrelationships among us all. Surely there is suffered some severe economic loss (McCulloch and Philip
some sense in which individuals within a species are more 1972). Where economic hardship occurs with little
related than individuals across species. King and Wilson probability of improvement it is unlikely that an individual
(1975a) have reported that the overlap between human and will be able to provide adequately for himself and for others
chimpanzee DNA and amino acid sequences is 99% and that sharing his genes. Suicide is also common in the terminally ill,
the variation is even less within the human population. While the elderly, and those with severe physical disabilities
this may reflect differences in variability of structural and (Dorpat, Anderson, and Ripley 1968; McCulloch and Philip
regulatory genes (King and Wilson 1975b), it does suggest 1972; Weiss 1968). Such individuals may show low
that the situation may be more complex than we have probability of engaging in future reproduction and would be
assumed. Although an offspring physically inherits one-half of relatively unable to engage in productive activity promoting
each parent's genes, many genes do not differ between his the welfare of others sharing their genes. Furthermore,
parents and may also not vary substantially within a adaptive behavior such as proper eating, sleeping, sexual
population. Thus more than one-half of the individual's genes behavior, grooming, and working habits may break down
may be identical to his parents' when one is considering genes under chronic stress or depression (Beck 1967; deCatanzaro
in kind (see also Washburn 1978). Even employing and Gorzalka 1979); such conditions may illustrate how
conventional formulas and considering the most proximate self-promoting behavior breaks down in states that may
generations, gene sharing among naturally aggregated precede suicide.
individuals may be quite complex; this is illustrated by the Suicide appears to bear a remarkable relation to
careful calculations of genealogical relationships within reproductive success. Suicide is much more common among
Yanomano villages by Chagnon (1979). I do not wish to imply single and divorced individuals and those with unstable
that the totality of gene sharing among individuals should be marriages than among those with stable marriages (Dublin
indicative of commonality of interest, nor that more closely 1963; Durkheim 1951). Those with stable marriages would
related individuals should not share especial interests. Rather, probably be relatively likely to reproduce and able to
I am suggesting that we have not yet defined adequately the promote the welfare of offspring. Indeed, individuals with
point of zero relationship and thus of no commonality of children are among those least likely to commit suicide
interest. The situation may be exceedingly complicated if we (Breed 1966; Dublin and Bunzel 1933). Also, recent divorce
fully consider the parameters of population genetics (see also or difficulty with members of the opposite sex is often a
Wilson 1975, pp. 119-20). proximate antecedent of suicide (McCulloch and Philip
Assessment of the role of altruism in human suicide thus 1972).
must await further developments in the theory of gene Furthermore, as discussed above, social isolation is a good
sharing and altruism, as well as further data on the impact of predictor of suicide. While this may be due in part to
suicide on others. Tentatively, given the indications that voluntary social withdrawal during presuicidal depression,
suicide may have little or negative impact on kin and the there is considerable support for the notion that social
current weight of theory against group selection, it may be isolation can precipitate suicide (see Stengel 1973). Socially
best to turn our attention to other variables. isolated individuals may be less likely than other members of
the population to form successful social relationships enabling
IV: Evolutionary tolerance of suicide. The final model them to advance their genes. Not only would they be unable
presents a promising alternative to the second and third to bring their genes into future generations without stable
models in accounting for aspects of suicide not attributable to relationships to members of the opposite sex; without contact
culture. This model would explain suicide on the basis of with kin they would be incapable of behaving to improve
special ecological and social characteristics of those their inclusive fitness, advancing their genes as they exist in
relatives. Consistent with this is evidence that suicide is most
committing the act. These characteristics may make suicide
common in the developed countries, where the family unit
immune to the pressures of natural selection. has become relatively less stable and smaller, and thus where
Suicide is an act of desperation. It occurs in extreme individuals have less opportunity to promote the welfare of
circumstances (Lingens 1972), particularly when an relatives sharing their genes.
individual feels incapable of coping with his present

Another excellent predictor of suicide is age; suicide occur under limited environmental conditions. Genes whose
frequency increases as a direct function of age, particularly in expression depends on environmental factors are by no means
men (Diggory 1976; Linden and Breed 1976). Although some rare (Caspari 1967; Herskowitz 1973).
cases have been reported, suicide is extremely rare among If an individual's propensity to promote his survival varies
children, almost never occurring in children under ten and in parallel with selective pressures, as described above, this
infrequently between ten and fourteen (Diggory 1976; Otto variation must have some inherited, biological basis. Such a
1972; Shaffer 1974; Toolman 1968). Productive and basis could conceivably consist of physiologically mediated
reproductive capacity may not be established by these ages, emotional or motivational variables. There is growing
and many of the difficulties in adjustment experienced by evidence that much of what we commonly call "emotional"
children may be transitory. Suicide becomes more frequent in behavior is in part mediated by organized biochemical
adolescence and early adulthood (Jacobs 1971; Hendin 1976), systems acting upon and within the brain (Barchas et al. 1978;
which could relate to increasing differential productive and Levine 1976). These systems involve neurochemical,
reproductive success. With increasing age, individuals may be hormonal, and other neuroregulating substances which
less likely to engage in reproductive activity, and respond in characteristic manners to stressful, novel, or other
consequently less able to spread their genes. Furthermore, stimuli. Such evidence comes from pharmacological
with increasing age, some individuals may be less capable of manipulations that affect specific neuroregulatory systems
engaging in productive activity that promotes the welfare of and biochemical assay data that correlate with behavior. For
related individuals such as existing offspring. It is interesting example, injection of central catecholamine stimulants may
that there are sex differences in both the frequency of suicide induce subjective feelings of well-being or remove protracted
and its relation to age (Linden and Breed 1976). This sex states of depression, while catecholamine antagonists may
difference could conceivably reflect differences between the induce or exacerbate states of depression. However, suicide
sexes in reproductive strategies and nurturance of offspring may relate more to serotonin than catecholamines; a number
(see Daly and Wilson 1978). of postmortem studies of brain samples of human suicides
A low capacity to promote one's genes thus appears to indicate that 5-HIAA, a metabolite of serotonin, is
account well for all of the best predictors of suicide. I would substantially reduced in suicides as compared to controls (see
now like to discuss why this relationship might occur. review by Murphy, Campbell, Costa 1978). Indeed,
If an individual's present and future behavior is unlikely to reserpine, which depletes catecholamines and serotonin, has
change the status of his genes, there may be no ecological been reported to induce suicide (see Lester 1972; Weil-
pressures preventing his suicide. If an individual is unlikely to Maherbe 1972). Physiological systems of this sort, responding
reproduce, unable to support himself and his family to stress and environmental stimuli, and controlling
adequately, and unable to contribute to the welfare of other, behavioral predispositions, may provide the inherited
reproducing individuals sharing his genes, his death may not variable substrate required by this model.
affect the frequency of genes he carries. Suicide would One other set of data may have bearing upon this model
consequently not eliminate any genes from the gene pool that and also perhaps on the third (altruism) model. Several
were not already eliminated. Thus, under the limited investigators (see Farber 1977) have noted that suicide
ecological conditions in which suicide appears to occur, there appears to cluster in families. While this could be due to
may be no selective pressures to prevent it. Conversely, common learning experiences in such families, it may also
suicide may be rare in relatively well-adapted individuals indicate that genetic factors underlie suicide. Kallman (1949)
because their death would have an impact on the gene pool. compared suicide rates in twins, finding higher concordance
In its weakest form, this factor would mean that suicide for monozygotic than dizygotic pairs. While Kallman did not
that occurred for other reasons would not be influenced by believe that he had established a relationship between
selective pressures. Thus, for example, a suicide that occurred genetics and suicide, Lester (1968), reanalyzing these data
because of cultural influences or behavioral disorganization with more sophisticated statistical methods, concluded that
consequent to stress might not cause any impact upon the they did indicate such a relationship. Other evidence
gene pool. In a stronger form, this factor might mean that (Gershon et al. 1976) suggests that affective disorders, which
selective pressures have favored the development of a are often precursors of suicide, are probably influenced by
variable motivational constitution in members of the species. genetics. These data may suggest genetically produced
Such a constitution might favor behavioral efforts toward variability among individuals in the propensity to commit
adaptation and self-advancement when such advancement suicide. However, the model does not require a large amount
was possible, but not when such advancement was no longer of variability among individuals; rather, it suggests that a
possible. It is also conceivable that this factor would allow tendency for individuals to vary in self-preserving and self-
cultural pressure or species benefit (perhaps mediated by promoting behavior is widespread in the population.
culture more than genes) to influence behavior so that it The final model suggests, then, that suicide is most
operated against one's survival when one consumed resources common when an individual is incapable of promoting his
without being productive. Where there is "no reason to live, " genes. This situation may mean that suicide is immune to
any otherwise trivial reason not to live might influence selective pressures in that it would not affect the gene pool.
behavior.
This model may require that selective pressures act
differentially on the same individual's behavior under Conclusions and research suggestions
different circumstances. When an individual is well-adapted,
selective pressures act against suicidal behavior. If the same These models are not necessarily mutually incompatible, nor
individual subsequently loses fitness and the capacity to do they necessarily exhaust all possibilities. Each model
further his genes, as might be the case, for example, in old age explains some, but not all, aspects of suicide, which perhaps
or severe illness, these selective pressures would no longer indicates that some combination of factors is responsible.
operate. Because of this, what might be favored in evolution is Suicidal acts are complex and diverse, and there is variability
the capacity to vary in the tendency to protect one's existence. in the documented proximate etiological factors. Different
It is not unusual for natural selection to favor behavioral factors may contribute to different cases, and it is likely that
predispositions that act differentially in different an interaction of several factors is necessary in any case. The
environments. Sexual, courtship, aggressive, hibernation, validity of these models remains to be established by future
migration, and other behavioral patterns of many species all research, which will have to determine the relative contribu-

Commentary/deCatanzaro: Biology of suicide
tion of the factors they suggest. The study of human suicide data on physiological correlates of suicide. Information about
also presents many problems that are not encountered in the involvement of group and kin selection may come from
other research endeavors; good discussions of these problems analyses of the social situations of suicidal individuals.
can be found elsewhere (e.g. Neuringer 1962). Hypotheses regarding evolutionary processes can be exam-
There is a substantial need for work reconciling the find- ined through logical analysis and computer modeling tech-
ings of human psychology and sociology with evolutionary niques. One can analyze the plausibility of hypothetical
theory. Any science of human behavior must be integrated evolutionary processes by simulating them mathematically.
with other developing areas of science. This is especially true Furthermore, hypotheses about ongoing evolutionary pro-
with respect to biological science, since behavioral processes cesses are not untestable in any absolute sense: we may have
are a function of biological organisms. Behavior is a means by difficulty devising means of testing them directly because of
which organisms adapt to their environments, much as physi- limitations on the length of time any one researcher can
ological processes act toward this same end. If we accept that sample, but, over generations, a sufficiently large sample of
organisms are subject to natural selection, we must accept that information may be gathered to test the validity of such
their behavior is also a function of such selection. hypotheses.
It is common practice to analyze animal behavior in terms Further advances in many areas of research, then, could
of its potential biological adaptive value. However, this kind contribute to the validation of the heuristic models presented
of analysis has not been practiced in psychological and here. Perhaps the major thrust of future work should involve
sociological approaches to human behavior, perhaps because careful investigation of the ontogeny of suicide and the classes
of artificial academic disciplinary boundaries. There is no of individuals displaying this behavior. Such work would not
reason why human behavior cannot be subjected to the be entirely new, of course. However, in classifying etiological
scrutiny of evolutionary reasoning just as animal behavior and factors and classes of people we musb have some conceptual
physiology are. Human behavior is as much a product of framework that orders the information we collect. That
biological variables as is the behavior of other organisms. conceptual framework may be best developed by formulating
Although learning and cultural variables produce much of the theories about the biological roots and ultimate causes of
topography of human behavior, we must still consider the suicide. There is a substantial need for general theories that
ways in which culture and learning interact with biological account for the many disparate findings and that help to
constraints. generate research hypotheses. Any general theory of suicide
Many may be concerned with the testability of hypotheses must account for the biological anomaly that this behavior
about evolutionary effects on behavior. Hypotheses about presents.
evolutionary history or the relations between selective pres-
sures and behavior in numerous previous generations may
indeed be difficult to test. Presumably, very little trace is left
of suicidal events that occurred one thousand or one million
years ago. Nevertheless, information about current suicidal Open Peer Commentary
acts and their proximate determinants is readily available.
Through systematic examination of this information we may Commentaries submitted by the qualified professional readership of
gain insight into the ultimate determinants of suicide. this journal will be considered for publication in a later issue as
For example, it may be difficult to obtain data directly Continuing Commentary on this article.
testing the proposition that "suicide has been tolerated by
evolution only when it has occurred in individuals who have by Hymie Anisman
had little capacity to promote their genes." This statement Department of Psychology, Carleton University. Ottawa, Ont., Canada K1S5B6
may or may not be true, but it is difficult to test its implica-
tions with respect to selection processes occurring over several Depression and suicide:
human generations. Many of its corollaries, however, are stress as a precipitating factor
readily tested. One corollary is that "suicide is currently most As indicated by deCatanzaro, the mechanisms underlying suicidal
likely in those with little capacity to promote their genes." behavior cannot readily be subsumed within a single theoretical
This statement is testable through the careful collection of framework. Among other things, the antecedent conditions associated
case histories, including data on the social situations of those with suicide vary considerably (depression, psychosis, chronic drug
committing suicide. The corollary, "suicide is very unlikely in abuse, terminal illness, and maintaining honor, as well as "(or king and
nonreproducing individuals who can support the welfare of country"). Not unexpectedly, attitudes toward suicide vary among
reproducing relatives," is similarly quite testable. By gather- cultures and may be dependent on the conditions precipitating the
ing together such information we can piece together a picture suicide (witness the condemnation of the misanthrope and the worship
that may begin to explain suicide. However, without more of the martyr).
general hypotheses, we may be unable to generate more
The proposition that suicide is related to the individual's capacity to
readily testable corollary hypotheses, and our power to
promote his genes (reproductive success) is not one that can be
explain may be considerably reduced.
readily assessed through experimental methods, despite the provi-
The situation with the other hypotheses presented is simi- sional suggestions for research advanced by deCatanzaro. Moreover,
lar. We can gain a substantial amount of information about evaluation of the evidence reviewed by deCatanzaro suggests that
the involvement of learning in suicide by collecting informa- attempts to relate gene propagation to suicide are gratuitous. In
tion on the backgrounds of those exhibiting this behavior. particular, a series of social conditions associated with suicide is
Cross-cultural studies may be especially valuable. The described (e.g., separation, divorce, isolation), and it is suggested that
involvement of genetics is also not impossible to factor out. each of these represents an instance where the probability of gene
Studies probing the frequency of suicidal acts in related propagation is reduced. However, each of these examples also
individuals raised in separate environments would be particu- represents an instance where the individual is confronted by uncontrol-
larly fruitful; the application of these and other behavior- lable stress or an instance where the individual does not have access
genetic techniques has provided important information about to social factors that might help in buffering aversive insults. Thus, the
genetic involvement in other kinds of behavior (e.g. Gershon source of suicidal intent can be derived from more conventional
et al. 1976). Further study of depression and suicide elicited models than that offered by deCatanzaro. Put simply, upon exposure
by pharmacological agents may shed light on possible physio- to aversive stimuli, behavioral and physiological changes occur in
logical and genetic determinants, as would additional autopsy order to meet environmental demands. If the stress cannot be dealt

with effectively through behavioral means, then the burden of coping a range of events from participation in war through direct self-
rests principally on physiological mechanisms. Under such conditions, destruction "for cause" such as terminal disease, to acts associated
use of central neurotransmitters, such as norepinephrine and seroto- with depression or other mental illness. It seems obvious that separate
nin, will increase and may exceed the rates at which these amines are consideration of these types of events is likely to be more fruitful than
synthesized. As a result, a net decline of these neurotransmitters may lumping them together.
be expected (see reviews in Anisman 1978; Stone 1975; Welch and Depression- and mental-illness-related suicide, for example, proba-
Welch 1970). Interestingly, if the stress is sufficiently protracted, bly represent malfunction of a set of brain systems of which the normal
physiological accommodation may occur in the form of increased functioning is extremely adaptive. This view is quite different from
activity of synthetic enzymes, thereby resulting in amine levels deCatanzaro's "pathology" category, which involves the failure of a
approaching control values (Kvetnansky, Mitro, Palkovits, Brownstein, system to function adaptively because of poor fit with unusual environ-
Torda, Vigas, and Mikulaj 1976; Weiss, Glazer, and Pohorecky 1976). mental circumstances. The point is that significant numbers of suicides
The initial stress-induced depletion of brain amines, and failure of involve endogenous depression not primarily related to environmental
amine levels to increase with protracted stress, may result in clinical stress. There is increasing evidence of neurochemical pathology for
depression, which in turn may precipitate suicide. such individuals as well as familial incidence rates suggesting that
The available data largely support a stress model such as that genetic factors (Pare and Mack 1971) are in large part responsible for
described here. Stress has been considered as a precipitator of the specific brain-chemistry changes associated with endogenous
depression, although it cannot be denied that stress may simply depression. The problem, then, is not fit with a changed environment,
increase the incidence of hospitalization of already depressed individu- but malfunction with deleterious effects even in very favorable environ-
ments. The fact that such cases represent only about 30% of suicides
als (see discussions in Akiskal and McKinney 1973, 1975). As
does not indicate that malfunction is a poor explanation of suicide: it is
indicated earlier, stress will reliably induce depletion of brain mono-
a very good explanation for one specific type of suicide.
amines, provided that the stress is one over which control is not
possible (Anisman, Pizzino, and Sklar 1980; Weiss et al. 1976). Of The remaining types are more complex. For many of them deCatan-
particular relevance to the present discussion, the amine depletions zaro's "pathology" model is probably the most relevant: human
are also dependent on the age of the animal (Ritter and Pelzer 1978), capacities for memory and reasoning are normally adaptive, as is the
as well as on social housing conditions (Modigh 1976; Welch and ability to feel and respond to pain, and a tendency to predict and
Welch 1970). It will be recalled that such factors are also associated anticipate future events. When all of these traits combine to present an
with the incidence of depression and suicide (see deCatanzaro). individual with a very dim view of his future prospects, the result may be
Finally, considerable data are available indicating that the monoamines suicide. In such cases (e.g. terminal illness, profound social disruption,
most affected by stress are those that underlie clinical depression or ostracism) the future reproductive potential of the individual may be
(Murphy, Campbell, and Costa 1978; Schildkraut 1978). To be sure, limited in any event, and his inclusive fitness only minimally affected by
inconsistent and contradictory evidence has been reported concerning suicide. However, even if every such suicide did result in a drop in
the role of any one neurotransmitter in depression. Nevertheless, it inclusive fitness, the contributory processes (e.g. memory) are still so
does appear that both norepinephrine and serotonin, and possibly also adaptive that they might be expected to be relatively little influenced by
their combined consequence in certain situations.
dopamine and acetylcholine, are intricately related to the depressive
symptomatology. There may, however, be one area in which human cognitive abilities
In considering a stress-depression model it should be remembered influence suicide (or phenomena involving hopelessness in general).
that the response to stress may vary widely between individuals. People do not recall or anticipate pain with anything like the vividness
Accordingly, apparently equivalent stresses may lead to pathology in of their impressions of past or anticipated joy. Might this not represent
one individual but have little effect on a second individual. These the result of negative selection against those individuals who have
individual differences may be due to any number of experiential, been able accurately to represent to themselves their own future
environmental, and organismic variations, and a role for genetic prospects for pain and suffering? Like all evolutionary questions, this
factors is certainly not being dismissed in the analysis of depression. It one is practically untestable at present, but it would be interesting to
has been reported that concordance rates for depression are higher in investigate some possible mechanisms involved in memory distortions
monozygotic than in dizygotic twins, even if the twins were reared or painful and pleasant events.
apart. At the same time the concordance rates do not approach values
that would preclude an environmental contribution to the affective by Edward G. Carr
illness (see review in Winokur 1978). It is likely that genetic factors Department of Psychology. State University of New York at Stony Brook. Stony
predispose individuals to depression provided that certain environmen- Brook. N.Y. 11794
tal conditions are present. In a recent review of the literature Gershon
(1978) indicated that catechol-O-methyl transferase, an enzyme Suicide: comments on deCatanzaro's
involved in catecholamine degradation, may serve as a genetic marker diathesis-stress model
for depression. These data support the hypothesis that catechol- The conceptualization of suicide that deCatanzaro argues most
amines (or associated enzymes) are linked to depression and may be strongly in favor of can be characterized as a diathesis-stress model.
genetically transmitted. However, this cannot be taken to support the The diathesis is polygenic in nature; the stress consists of any of a
contention that depression or suicide is related to differential reproduc- number of common but serious life problems.
tive success. At this stage in the analysis of factors that promote The first question is how to test this model. DeCatanzaro suggests
suicide it is not advantageous to rely on a sociobiological model such that one major corollary, namely that "suicide is currently most likely in
as that presented by deCatanzaro. those with little capacity to promote their genes," is easily tested by
examining case histories. However, he forgets that this "corollary" is
by D. Caroline Blanchard really nothing more than a summary statement based on a series of
Bekesy Laboratory of Neurobiology. University of Hawaii. Honolulu, Hawaii 96822 studies conducted by other investigators who examined case histories
involving the terminally ill, the elderly, alcoholics, those involved in
Biological variation and suicide disrupted social relationships, and the like. Thus, there is an inherent
Complex phenomena such as cancer or aggression - or suicide - may tautology here. One cannot test a model by relying upon the same set
involve a number of diverse processes with different antecedents and of data used to construct the model in the first place. In contrast, the
different evolutionary outcomes. The functional or adaptive signifi- genetic aspect of the model can be tested in a noncircular fashion.
cance of the component processes treated under one rubric may vary Behavior-genetic techniques have proven useful in understanding
tremendously, and progress in understanding them depends largely on schizophrenia and certain adult affective disorders. Perhaps the same
an adequate analysis of the different events involved. methods of analysis could be employed to study suicide, as deCatan-
OeCatanzaro's "suicide" is not defined, but examples given suggest zaro himself suggests. In this regard, one would think that studies of

Commentary / deCatanzaro: Biology of suicide
monozygotic and dizygotic twins reared together and apart could them the moth story. That is my inclination in the case of suicide.
provide the most critical tests. There is another, and rather less obvious, way in which domestica-
The second question concerns what clinical advantages accrue tion can enter into the argument. Suicide, even in a wild population,
from adopting a model with a strong biological component. Presum- requires a Darwinian explanation only if there is some genetic contribu-
ably, an important reason for developing models of suicide is to help tion to variation in the tendency to commit suicide. This is so
do something about the problem. In this respect, the primary clinical fundamental, by the way, that it is rather surprising that deCatanzaro
advantage of knowing whether there are individuals who possess waited until near the end of his article before mentioning it. Let us
genes that facilitate suicidal behavior would be in terms of prevention. accept, for the sake of argument, that there is such genetic variation in
That is, such individuals, once identified, could be taught how to cope human populations. Then, "gene for suicide" means a gene that,
with environmental stressors that would normally interact with their compared with its alternatives (alleles) at the same chromosomal
genetic endowment to increase the probability of suicide. Once they locus, makes individuals more likely to kill themselves given a specified
acquired such coping skills, they should be less likely to respond to environment. It follows that a gene for X in environment A may turn out
stress in a suicidal manner. to be a gene for Y in environment B. Ridley and I have already made
In the absence of a systematic body of data demonstrating an this point with respect to homosexuality (Ridley and Dawkins in press).
important role for genetic factors in suicidal behavior, it would be a A gene for suicide in the environment of modern western society may
serious mistake to underestimate the impact of social learning vari- have had some completely different effect in Pleistocene Africa. That
ables in the etiology and maintenance of suicidal and subsuicidal effect could have been almost literally anything, for development is
behaviors. At present, a sociobiological approach to the problem, such a complex process. It does not have to bear any resemblance to
while intriguing, appears at best to offer the possibility of understanding suicide; a modern gene "for" suicide might, in the Pleistocene, have
the subtle modulating effects that result from constitutional factors. At been a gene "for" resistance to some disease. Nearly all mutations
worst, such an approach may lead researchers to overemphasize have many effects anyway: the outwardly visible behavioural effect that
biological variables at the price of ignoring coping skill deficits and interests us may be the tip of an internal biochemical iceberg. If a
powerful modeling and social reinforcement factors whose role in the biochemically advantageous gene has disadvantageous behavioural
origin and control of suicidal behavior has been clearly documented. side-effects, selection may work on other "modifier" genes to change
the side-effects. But if the side-effects only manifest themselves in
modern "civilized" environments, we cannot expect the necessary
by Richard Dawkins modifiers to have been selected.
Department of Zoology. University of Oxford. Oxford OX 1 3PS, England Turning to deCatanzaro's hypothesis that suicide might be favoured
by kin selection, my principal feeling is that the domestic animal
Domesticity, senescence, and suicide
hypothesis is much more likely. But I would like to correct one rather
DeCatanzaro's introductory section "Evolution and behavior" is sensi- important error in his treatment of the theory of kin selection. "The
ble and correctly forestalls some of the misconceptions to which social coefficients of relationship most commonly used . . . give a brother or
scientist readers might be prone. I do, however, have some critical parent one-half of an individual's genes, a cousin one-eighth, and so
remarks to make about his treatment of suicide itself. on. . . . Surely there is some sense in which individuals within a species
I am not qualified to judge his review of the sociological literature on are more related than individuals across species . . . the overlap
suicide, so I will confine myself to two miscellaneous points. The first is between human and chimpanzee DNA and amino acid sequences is
a minor statistical worry. What can it possibly mean to say that suicide 99% . . . and the sequences vary even less within the human popula-
"ranked as the ninth most frequent cause of death in the United States tion. . .. Thus more than one-half of the individual's genes may be
in 1975," and that, in young adults in many countries "it is the third or identical to his parents' when one is considering genes in kind."
fourth most frequent cause of death"? Obviously everything depends DeCatanzaro here falls for what I have elsewhere criticised as "Wash-
on how the other causes of death are classified. Is "disease" one burn's Fallacy" -misunderstanding number 5 of my "Twelve Misun-
cause of death, or are "cancer" and "heart failure" classified sepa- derstandings of Kin Selection" (Dawkins 1979). Let me briefly explain.
rately? Depending wholly on how finely subdivided the other causes of It is, of course, perfectly true that all of us share more than 90% of our
death are, we may be very impressed, or very unimpressed, with the genes. The calculations of coefficient of relationship take all that for
importance of suicide as a cause of death. My second remark about granted and refer to the extra relatedness of close relatives over and
the human evidence is more constructive. It is that it would have been above the commonly shared genes. DeCatanzaro is aware of this, but
very interesting to have been told something about the incidence of he doesn't understand why. His intuitive feeling is that altruism towards
suicide in bushmen, or other tribes who might be supposed to be living nonkin may be favoured by selection, because even nonkin share the
in an environment similar to the one in which they have been subjected vast majority of their genes. But let us express this intuitive feeling in
to natural selection. This brings me to the part of the article that I am terms of a simple genetic model and see where it leads. Imagine a
gene, U, for universal altruism. Possessors of U are likely to commit
qualified to criticize.
suicide for the benefit of any members of their own species. U is shared
Of deCatanzaro's four models, the first two do not seem to me to be
by 99% of the population, so at first sight it seems that U is, in effect,
clearly distinguished. I would lump them together as the "domestic
caring for copies of itself, in just the same way as genes for kin altruism
animal hypothesis." A domestic animal lives in an environment other
are supposed to. But now look at the 1% unaccounted for. Suppose
than that in which its genes were naturally selected. Domestic pigs are
that this minority possesses a gene K for kin altruism. K individuals
obviously hopelessly ill-suited to avoid predators, and their behaviour
sacrifice themselves only for close kin. It is now easy to see that the K
in the presence of a wolf might well look like suicide to us. It would not
gene is bound to spread through the population at the expense of the
surprise us, however, because we suspect that the wild ancestors of
U gene. This is because the U gene, by its indiscriminate altruism,
modern pigs would have behaved very differently. As R. D. Alexander benefits its rival, the K gene, every bit as much as it benefits copies of
has observed, Darwinians do not worry unduly about moths' habit of itself. The K gene, on the other hand, since it causes individuals to
immolating themselves in candle flames. Candles, indeed noncelestial discriminate in favour of close kin, statistically benefits copies of itself
light sources generally, are recent innovations in the world of moths, at the expense of its U rival. I hope this very crude model will give at
and we are prepared to accept that natural selection due to candles least some intuitive understanding of the nature of deCatanzaro's
has not yet had time to work on their gene pools. The very sensible error.
orienting technique of maintaining a fixed compass bearing towards
light rays coming from infinity can become suicidal if the moth tries to Finally, deCatanzaro's fourth model does not, it seems to me,
do the same thing to light rays radiating out from a candle: it describes deserve the preference that he gives it. He suggests that suicide is
a neat logarithmic spiral into the flame. My first impulse whenever most likely to occur among people who are relatively unlikely to
somebody bring up some odd aspect of modern human behaviour and reproduce anyway, for example the old, the sick, and the divorced. No
defies me to "explain that with your selfish genes if you can," is to tell doubt many critics will object that these categories of people have

Commentary/'deCatanzaro: Biology of suicide
more obvious reasons for a statistical tendency to suicide than evil. . . . Self-existence, or the suicide that terminates it, is not the
concern for their genes. It is not clear to me that such common-sense central question of philosophy. The hypothalamic-limbic complex
proximal explanations are necessarily more parsimonious than deCa- automatically denies such logical reduction by countering it with
tanzaro's type of functional explanation, but in any case I want to raise feelings of guilt and altruism (Wilson 1975, p. 3.)
a different kind of problem. "If an individual's present and future Thus begins Edward Wilson's famous work Sociobiology. Wilson
behavior is unlikely to change the status of his genes, there may be no appears to deny the seriousness of the question of suicide, and,
ecological pressures preventing his suicide. . . . Suicide would conse- perhaps, even the possibility of suicide, by arguing that the emotions
quently not eliminate any genes from the gene pool that were not flooding consciousness will determine what the individual does and
already eliminated." DeCatanzaro might be right if it were literally true that the genes ultimately determine those emotional inputs. Of course,
that suicide was confined to individuals who have not the slightest hope Wilson later greatly modified this assertion of biological determinism,
of ever reproducing. Clearly, however, he is only talking about a especially in On Human Nature (1978). But, while he did attempt to
statistical tendency. The genetic impact of death in an elderly bachelor, explain one seemingly unadaptive form of behavior, homosexuality, he
say, is less than the genetic impact of death in a newly married young did not return to the question of suicide. This is most unfortunate, since
man, and selection would penalize a tendency to suicide in the young suicide not only poses a great challenge to the theory of evolution, as
more than in the old. But consider the options open to an elderly deCatanzaro makes so clear, but also, being far more restricted to
bachelor. It may be true that he is quantitatively unlikely to reproduce human beings than homosexuality, will ultimately reveal far more to us
anyway, but he is still even less likely to reproduce if he kills himself about human nature. Suicide has special bearing on the question of
than if he soldiers on. DeCatanzaro's argument is incomplete. His biological and cultural determinism versus individual freedom.
suggestion of a quantitative weakening of selection against suicide in DeCatanzaro has certainly done a masterful job of bringing a mass
certain classes of people needs to be combined with some positive of interdisciplinary research and theory to bear on suicide and relating
reason for suicide. He does, indeed, suggest that his fourth model this to basic questions about the theory of evolution. He has shown
might be combined with one of the other three. beautiful restraint and interdisciplinary balance in evaluating the many
Let me suggest another, and I think rather fruitful, way in which he complex factors that seem to be involved in the many complex forms
might develop his fourth hypothesis. It is that he should combine it with of suicide. But I believe there are two ways in which he could
the Medawar-Williams theory of senescence (Williams 1957). Any new strengthen his argument.
mutation, as we have seen, is likely to have many different effects First, there is the vital question of the validity and reliability of the
("pleiotropy"). These effects will not necessarily make themselves felt data. It is all too clear that there are very broad statistical biases in the
at the same time in life. Some will affect the individual in early official statistics on suicide (Douglas 1967). There are now at least
embryology, others in childhood, others in young adulthood, others in three comparative studies of medical examiner and coroner proce-
middle age, others in old age. It is rare for any one of a mutation's dures for certifying the causes of death that reveal how these biases
effects to be advantageous, and the chance of two of its effects being come about. (See, for example, Douglas 1971, pp. 79-132.) Some of
advantageous is negligible. If a mutation has a beneficial effect at one the works deCatanzaro draws on rely heavily on official information.
particular time in the individual's life history, its effects at other times of Fortunately, whether by happenstance or by design, some of the major
life are likely to be disadvantageous for purely statistical reasons. A biases in the official statistics undercount cases in the direction
mutation that has a bad effect in youth will tend not to persist in the supportive of his major arguments. Any experience in the medical
gene pool, however beneficial it may be in old age. A gene that has a world quickly reveals that the intentional killing of oneself (suicide), and
beneficial effect in youth, on the other hand, will tend to be passed on intentional cooperation in "pulling the plug" (ending medical support),
to future generations, however lethal its effects in old age. This is, very are far more common among terminally ill patients than is indicated by
briefly, the Medawar-Williams theory: senile decay is due to the official statistics. Doctors, nurses, and others cooperate tacitly in not
accumulation in the gene pool of lethal or sublethal genes that have not noticing, hence hiding, these suicides. Since most of these patients are
been removed by natural selection because of beneficial early effects, older and are all, according to medical prognosis, doomed to die soon,
and because their deleterious effects do not make themselves felt until their deaths fit the model of "evolutionary tolerance." Some of these
after the individual has had time to reproduce and pass them on. deaths also involve altruism because they remove a burden from the
If there is a genetic tendency to suicide (a big "if," but there has to families.
be if deCatanzaro's paper has any point at all), the genes concerned DeCatanzaro could also have strengthened his argument greatly by
are properly described as lethal, albeit they work via complex behav- carefully considering the arguments of Jean Baechler in Suicides
ioural routes. Their presence in the gene pool could be explained in the (English translation 1979; originally published 1975). Starting with the
same way that the Medawar-Williams theory explains the presence of conclusion that official information on suicide is often invalid, generally
other lethal genes. I find this a more complete explanation of deCatan- highly unreliable, and always very shallow, Baechler carried out the
zaro's observations on statistical biases in suicide statistics than his most extensive investigation of case studies yet accomplished, both
fourth hypothesis on its own. But my main reaction to the paper is the within Western cultures and cross-culturally. His study is interdisciplin-
one I gave earlier, that the whole problem he raises may be a ary in the same creative spirit as that of deCatanzaro. But Baechler
nonproblem. Darwinians need worry about suicide only if it is under was able to propose a systematic theory of suicide that goes beyond
genetic control, and then only if it is widespread in wild animals specifying the possible contributing factors, the level of theorizing that
(humans), observed in the environment in which their genes were deCatanzaro feels constrained to remain at because of the more
naturally selected. limited data with which he was dealing.
Though the details are far too complex to present here, Baechler
has summarized his theory very clearly. He argues, as do most
by Jack D. Douglas theorists of human nature today (e.g., Wilson 1978), that any adequate
Department ot Sociology. University ol California, San Diego. La Jolla, Calif. 92093 theory of behavior must take into consideration the initial genetic inputs
(human nature); the lifetime of experience and learning that build upon
Baechler's theory of suicide the genetic inputs but are by genetic programming partially autono-
Camus said that the only serious philosophical question is suicide. mous; the relatively stable personality tendencies to react in specific
That is wrong even in the strict sense intended. The biologist, who is ways to specific situations that result from the interactions of the
concerned with questions of physiology and evolutionary history, genetic inputs with experience; and the immediate (concrete) situa-
realizes that self-knowledge is constrained and shaped by the tions that people encounter in their lives. (One part of the experience is
emotional control centers in the hypothalamus and limbic system of the so-called culture variable, but it is vital to see that "culture" is
the brain. These centers flood our consciousness with all the important only insofar as it directly affects the experience of individu-
emotions - hate, love, guilt, fear, and others - that are consulted by als. In all civilizations cultural experience is quite pluralistic, so that
ethical philosophers who wish to intuit the standards of good and different individuals may experience different parts of the culture.)

Baechler, unlike most social scientists, but very like geneticists and tion that suicide could in fact be evolutionary based, I shall try to react
behavioral biologists, recognizes that "life is a struggle" (or full of to the models described. With regard to the first model, that of suicide
conflict). Each individual, presumably as a result of genetic program- as a learned behavior, it seems that the evidence presented actually
ming, tries to protect himself in the struggle, and allow himself to argues against the hereditary component. The presence of suicidal
continue actively pursuing the struggle, by building personal "security" behavior in humans and not other animals and the fact that "much of
for the self. This is done primarily by using two basic strategies, which the available evidence concerning suicide can be construed as support
are both social strategies because of the fundamental importance of for a model based on learning and cultural evolution" seem to preclude
social relations in human life. These are the strategies of power over a strong genetic component. Further, such findings as that suicide is
others and dependency upon others. Most people use a combination relatively infrequent in Catholic countries, for example, certainly
of these two because this is the optimal strategy for building personal suggest that cultural and learning factors are most important - unless,
security, but those thwarted in one may pursue the other almost of course, Catholics have in some way selected genetically for
exclusively, thus making themselves more vulnerable to emotional nonsuicidal genes.
insecurity from defeats in the struggles of life. Pursuit of either or both The learning theory model is not the only one with which I had
strategies may in itself demand that one risk death by resorting to trouble. I also had some difficulty with the pathology model and the
suicidal actions in certain situations. (This would, for example, be true altruism approach. My main question with regard to the former deals
when a suicidal gesture used to get revenge, which enhances one's with whether or not the capacity to cope must, as the author seems to
power, leads to death.) But the use of these strategies may also fail. imply, be inherited; cannot coping methods also be learned? As for the
(For example, one may be rejected in love-dependency.) When this altruism model, it strikes me that the author is confounding what is with
happens in extreme ways, thus producing extreme feelings of rejection regard to isolation with what an individual feels with regard to isolation.
or failure, then individuals respond with varying degrees of depression The net result is ascription of motive based upon outcome rather than
and hopelessness, presumably because of genetic and personality upon actual experience.
differences (which may be amplified or damped by cultural experi- While a longer commentary would afford me the opportunity to react
ence). It is only in extreme situations that "one can be driven to the further to the three models I have very briefly mentioned, in the
point where one can no longer act and seize on the possibility of remainder of my comments I would prefer to focus on the model
leaving the fight" (Baechler 1979, p. 455). dealing with evolutionary tolerance. Here deCatanzaro states that
We must note that the individual may "seize" the possibility. We suicide is likely to be selected for in individuals who experience
must also note that the individual may or may not choose to risk his life difficulties, become nonproductive, and see little hope for improvement
by pursuing his strategies to extreme positions. For, ultimately, after all in the future. It is noted that persons incapable of reproducing or
of the powerful predisposing factors are considered, human beings do otherwise promoting their genes are those in whom suicide would be
have some freedom of individual choice. "The conclusion is unambigu- evolutionary "tolerated." As support for this contention the author
ously imposed: suicides, in all their different meanings, are neither cites data indicating that suicide is more common among single and
perversions nor scandals. They are coeval with the human condition, divorced persons and those with unstable marriages. Also cited as
for human life is an endless struggle where each of us by his own free "predictors" of suicide are social isolation and increasing age. All of
will risks his life each moment" (Baechler 1979, p. 451). these factors are seen as "signs" of decreased chances of promoting
There will always remain some element of individual mystery to one's genes. For example, it is stated that divorce and suicide are
suicides, as there will to all human action, because man is "con- related, and it is thus concluded that suicide is tolerated in the person
demned" by genetic programming to make his own choices in the full who, once divorced, has suffered a reduction in probability of repro-
light of the genetically determined emotional inputs from the duction. A more parsimonious and probably more easily supported
hypothalamic-limbic system, the situations he faces, and his aspira- explanation, however, might be that suicide and divorce are both the
tions. Man is predetermined biologically to be both constrained and result of faulty interpersonal abilities and maladaptive learning. Like-
free. Suicides are the outcome of profound personal struggles and the
wise, the other "signs" of decreased reproductivity, such as unstable
ultimate choices that are made to deal with those struggles.
marriage, nonmarriage and isolation, may also be related to suicide
only by virtue of the fact that all of them may stem from basic
psychological dysfunctions. DeCatanzaro's conclusions concerning
by Marshall P. Duke
the relations between suicide and signs of reduced ability to promote
Department of Psychology. Emory University. Atlanta. Ga. 30322
the genes may be the result of a common logical error: b and c occur
Feasting on the sociobiology of suicide: somehow I concurrently; what is their relationship? Conclusion I: b causes c
still feel hungry . . . (divorce causes suicide); or conclusion II: a causes b (a representing
basic interpersonal dysfunction); a causes c; thus b and c occur
While I applaud deCatanzaro's efforts to delineate a biological together (faulty relating causes divorce; faulty relating causes isolation
perspective on suicide, I came away from the article with a mixture of and depression).
puzzlement and disappointment. I almost feel as if I had been prepared In addition to making this logical mistake, the author seems to leap
for a sumptuous feast and the waiter stopped serving the food just after from prediction of suicide to precipitation of suicide. On p. 270 for
the appetizer. I was intrigued by the author's goal of establishing the example, it is first stated that "social isolation is a good predictor of
adaptational-biological-evolutionary impact upon suicide and looked suicide." One sentence later the author writes, "social isolation can
forward to the presentation of the models that were discussed. I am precipitate suicide." Again, I am left with the question of whether social
assuming that other important psychological and sociological models isolation causes suicide or whether suicide and social isolation are
were deleted in the interests of exposition, so after simply registering both caused by something else again.
my concern that several viable and empirically supported models were Another problematic implication of the evolutionary tolerance
not discussed, I will proceed with some reactions to the materials perspective runs counter to data cited by the author. It is correctly
actually included in the target article. noted that suicide occurs more often in males than in females,
A problem I had initially was accepting deCatanzaro's assumption regardless of age, yet males typically continue to be able to reproduce
that "the innate component [of biological development] should reflect far beyond the average age for female menopause. If being unfit to
countless generations of selection for adaptiveness, and we should reproduce removes the negative pressure against suicide, one would
thus be able to account for most [!] behavior in terms of evolutionary expect postmenopausal females to be more suicidal than men ot
contingencies that affect it." Such a statement cries out for an equivalent age - statistically this is simply not the case.
empirical foundation upon which to rest. If I, as a reader, cannot accept While my comments to this point have been theoretical and logical,
this assumption (and I'm not sure I can) the worth of the remaining my final reactions to the article rest on some scholarly concerns. The
effort put forth by the author is in question. extensive use of secondary sources seems to have laid the author
However, given a scholarly acceptance of deCatanzaro's assump- open to some criticism. For example, I was struck by the statement

that "indeed, reserpine, which depletes catecholamines and serotonin, is the same. Thus suicide can be placed squarely in the process of
has been reported to induce suicide." Having never felt that such a natural selection and evolution.
relationship had been clearly supported, I went to one of deCatanza- Model IV, by suggesting that sociologically defined groups with
ro's sources (Lester 1972) and found the following statement: relatively high suicide rates produce, on average, fewer offspring and
It is claimed from time to time that particular drugs produce depres- provide less favorable conditions for their genes, adds support to the
sion and suicidal tendencies in patients. Berger (1967) reviewed the notion that the elimination of this group from the gene pool is
evidence that rauwolfia serpentia (reserpine) induces suicidal behav- beneficial.
ior. It is difficult to evaluate the research since the patients given the One must, however, interpose a caveat. Although it is true that these
drug are often not matched for symptoms and degree of distur- groups have relatively high suicide rates, many suicides are not in
bance with patients given the placebo. Secondly, those who react these categories. Moreover, the old were once young and may earlier
suicidally to the drug may have pre-existing tendencies to depres- have produced large families. The divorced were once married and
sion. Berger concluded that the bulk of evidence available indicated reproductive. The isolated may well at some point have been parents.
that reserpine was more likely than any other central nervous system There exist many offspring of forebears who have committed suicide.
depressant to produce depression and lead to suicide. But this The contrast between the suicidal and the nonsuicidal with regard to
conclusion must be regarded as a tentative one. (Lester 1972, p. reproduction, at least, would appear to be somewhat overdrawn.
309) Nevertheless, there remains some force in the argument. Those
Indeed, Lester |q.v.| presents a very different picture of the reserpine- eliminated tend to provide neither good quality nor quantity of genetic
suicide relationship. material.
In what may be seen as a further scholarly lapse, in discussing the But there is another effect: the loss of valuable genes. Along with
possible relation between genetics and suicide deCatanzaro stated genes vulnerable to stress there may exist desirable gene combina-
that "Lester (1968), reanalyzing these (Kallman'sl data with more tions in an individual, even some for genius. If the suicide is young, he
sophisticated statistical methods, concluded that they did indicate may well not yet have reproduced. If older, his loss may impair the
such a relationship." In fact, Lester (1968) actually stated, "It must be welfare of the genes of others. On balance, however, one might
concluded, therefore, that it is not possible to dismiss inherited factors speculate that the net effect of the two countervailing forces is
in the determination of suicidal behavior on the basis of present weighted toward elimination of the undesirable. The infrequent loss of
evidence" (p. 320). There is quite a difference between concluding high-quality genes would not seem to outweigh the probably more
that a relationship is indicated and that one cannot yet be ruled out. frequent elimination of stress-susceptible ones. Not only can suicide be
I began by mentioning that the target article left me feeling unsatis- "tolerated" by evolution, as deCatanzaro suggests, but, on balance,
fied. Much of this feeling stemmed from the dearth of clear research from a purely evolutionary point of view and ignoring the personal grief
direction at the end of the article. I still don't know where we ought to and social loss, it may well entail a modest gain.
go; or whether the genetic approach is worth pursuing aggressively Overall, one can congratulate the author on having addressed
and enthusiastically. While the evolutionary argument is interesting, it himself incisively to a problem rarely considered by psychologists.
seems that most of the data support a minimal genetic component
in the act of suicide itself. On the basis of his review of suicide Lester
(1972) has stated: "It is unlikely that the tendency to suicide per se is by Gary Frieden
inherited, but it is possible that genetic factors play a role in the ASt Market Research. Inc.. Los Angeles. Calit. 90046
formation of personality types susceptible to suicide or in the incidence
of particular mental disorders which facilitiate the appearance of Self-destructive behavior: suicide, shocks, and worms
self-destructive behavior" (p. 25). DeCatanzaro's approach considering suicide in terms of four models,
In general, I was disappointed by many things in deCatanzaro's is an interesting one. I would like to expand on his points pertaining to
laudable attempt to establish an evolutionary perspective on suicidal learning, social, and cognitive factors. The purpose of this commentary
behavior. Had I not been distracted by a number of what seemed to be is to elevate the importance of these variables in any model of suicide.
forced arguments, illogical conclusions, and scholarly lapses, I might DeCatanzaro stated that self-destructive behavior may occur when
have come away with a more favorable view of the effort. organisms have not learned the negative contingencies surrounding
such acts. In a review of self-punitive behavior (Renner and Tinsley
1976), it has been shown that such maladaptive behavior extinguishes
by Maurice L. Farber when organisms become aware of the negative contingencies asso-
Department ot Psychology, University of Connecticut, Storrs, Conn. 06268 ciated with behaving in a self-punitive fashion. Similarly, learning
experiences can actually produce self-punitive behavior as well. For
Suicide as natural selection
example, one study (Stretch, Orloff, and Dalrymple 1968) has demon-
Clearly, as deCatanzaro's target article asserts, suicide presents a strated that an aversive stimulus, such as an electric shock, can be
theoretical paradox in that it appears to run counter to the process of reinforcing because it simultaneously serves as a discriminative stimu-
evolution. His model 1 considers the possibility that suicide could be lus for the absence of punishment: organisms first received avoidance
conceived as outside this process. But, if it can be demonstrated that training in that their responses could indefinitely delay an electric
suicide involves natural selection, it must then be viewed as being shock. In a second phase, shocks were presented on a fixed-interval
within the evolutionary process and contrary to the model. This schedule; nevertheless, the organisms' responses increased! Interest-
commentary will attempt to place suicide within the evolutionary ingly, the only way the animal could be cued as to the absence of
framework. shock in a subsequent interval was to cause the shock actually to
Although it is true that suicide operates against the survival of the occur. Thus, the punishment was really a discriminative stimulus for the
organism, it does so selectively. DeCatanzaro's model IV makes use of absence of punishment; as a result, it was reinforcing. A most intriguing
this fact, but, as will be shown, supports its conclusions with somewhat anomaly indeed - a punisher and a reinforcer at the same time.
flawed analysis. This aspect of learning relates to a point made by deCatanzaro. He
Suicidal behavior is determined not only by situational stresses, but stated that attempted and completed suicide differed in that there
also by characterological vulnerability to such stress (Farber 1968). could be secondary gains from attempted suicide (perhaps sympathy
This vulnerability would appear to involve both acquired and innate or better treatment from other people), but not from completed suicide.
components. Of special interest to us here are individuals with these Yet he later added that completed suicide could result in a removal of
innate vulnerabilities. In suicide, precisely such people are eliminated. pain or other aversive events - that is, negative reinforcement. It
Just as the slowest-running deer are eliminated by the pursuing wolves, seems, then, that there are gains to be derived from both forms of
so too are humans with vulnerable character structures eliminated by suicide, depending on the individual's needs at the time of the act.
environmental stress. Although it is done by their own hand, the effect There is another factor that can contribute to self-destructive

behavior; deCatanzaro referred to this as "complex and unusual Of course Skinner's (1969) argument was presented before kin
cognitive events." I would like to expand on this category by briefly selection had much of a chance to be developed theoretically or with
explaining a model of "role confusion" that I am currently developing. significant examples (as is now the case) for bird and mammal
Suppose people were to suddenly wonder who they really were. They species.
could become confused about whether a previous sample of their DeCatanzaro's discussion of the extent of suicide by animals can be
behavior represented merely role playing or their "real" self. At such extended. He notes that suicide has not been commonly identified by
times, they may wish to "test" themselves to find out more about ethologists in field situations and suggests the possibility that this may
themselves (and to answer the question, "Was that really me?"). This result from inadequate observations. In my opinion, however, we are
behavioral test would consist of behaviors similar to the acts that today not in a situation where this is a very probable explanation.
produced the confusion itself. Lo and behold, people depressed about Rather, the main message I derive from comparative animal behavior is
the possibility that they are worthless, incompetent, masochistic, that, in spite of the intensive study of thousands of animal species in
sadistic, and so on, would continue responding in that fashion in an nature, suicidelike behavior is rarely found, and where it occurs it
effort to resolve the role confusion. probably has its basis in parental behavior or kin selection. It is not
The above model was tested (Frieden 1978) by having respondents expressed by individuals who have direct residual reproductive poten-
role-play as if they enjoyed suffering - that is, as if they were masoch- tial, in striking contrast to most human suicide.
ists. Some people were then led to experience role confusion by Because, as deCatanzaro points out, suicide has been a little-
suggesting to them that their previous behavior may actually have treated subject in this context, any analysis will be incomplete. My own
reflected their true desires. Those who believed they had acted on their consideration identifies four patterns of self-destructive behavior:
own accord later chose to be bound, blindfolded, and verbally abused, Emigration. Animals may make one-way movements to areas
and to receive painful electric shocks for twenty minutes. Not only that, where they do not survive. Many animals move to and from alternative
but when later asked if they would be willing to eat a dead worm, 75% ranges. Such migrations, although they are hazardous and may involve
agreed! (No one ever experienced any of the above - once they made massive mortality, are clearly adaptive, involving movement to and
their decision, the study ended, and they were extensively debriefed.) from seasonally available resources and breeding grounds. To the
What does this tell us about suicide? It should make us aware that extent that emigrators or their descendants do not eventually return to
the "complex and unusual cognitive events" to which deCatanzaro the site of origin, they contribute nothing to the population of origin and
refers can be quite powerful indeed. Not wishing to play down the can therefore not influence the adaptive characteristics of the source
biological role, I am instead citing the importance of cognitive factors. population. But to be adaptive the return need not be immediate and
When people attribute causality to themselves for unfortunate events can involve the progeny of the emigrators. This is so, for example, in
that they did not in reality produce, the self-questioning process can be the case of the painted lady butterfly, Vanessa cardui, which resides in
maladaptive. In another study (Frieden and Fraser 1978), an inverse Mexico. Northward emigration from there in the summer involves a
relation was found between one's masochistic tendencies and one's wave of movement including three generations extending to Canada.
sense of self-worth - the more subjects enjoyed suffering, the worse The fall return mirrors the exodus, and the annual return to Mexico is in
they felt about themselves. Furthermore, those whose self-esteem was essence a multigeneration migration and therefore adaptive (Arthur
experimentally lowered in the laboratory subsequently chose to suffer Shapiro, personal communication).
voluntarily. It is not surprising, as deCatanzaro pointed out, that guilt The sometimes spectacular and apparently one-way movement of
feelings and negative self-concepts correlate with suicide (Kaplan and tropical butterflies and moths may not involve a return to the area of
Pokorny 1976). origin. If not, the adaptive explanation must resort to parental behavior
The learning, social, and cognitive factors, then, cannot and should or kin selection arguments involving relief of population pressure upon
not be viewed as having only a minor influence on suicidal behavior. progeny or other relatives living under submarginal conditions. The
Whether the gene pool is enhanced or unaffected by one's committing pattern in the case of a classical emigrator, the Norwegian lemming
suicide, the tragic act itself has multiple causes, all of which deserve {Lemus lemmus), appears to involve movement up and down slopes to
further study. areas of seasonally preferred habitat (Kalela 1962). For some individu-
als this may involve movement to places where they cannot survive
Acknowledgment (Clough 1965), but this is little different from the imperfection of
I wish to thank Scott C. Fraser, at the University of Southern California, for long-distance migration by birds, whales, and insects, which may
providing valuable assistance with the understanding and implementation of this sometimes include massive mortality. In the face of completely biohos-
research. tile conditions, Caribbean lizards may leap into the sea and swim and
float away with a high probability of death (Schoener, personal
communication). But the consequence of remaining in an environment
by William J. Hamilton III without adequate food is certain death, and movements of terrestrial
Division of Environmental Studies. University of California, Davis, Calif. 95616 lizards into the open sea can, in fact, be best identified as an optimal,
albeit high risk, strategy. Additional supportive evidence for this
Do nonhuman animals commit suicide? interpretation is that lizard species living in coastal environments and
on small islands are, in fact, better able to stay afloat for prolonged
DeCatanzaro's evaluation of the biological basis of suicide follows
intervals than interior mainland species (Schoener, personal communi-
logically from the current attempt to evaluate all aspects of human
cation). Thus, positive selection for swimming ability by a terrestrial
behavior in the context of comparative animal behavior and compara- animal is established.
tive analysis of diverse human societies. This is at least a strong
alternative to the classical psychiatrists' motivational theories. Completed suicide by social insects. The significance of self-
This approach is effective and needs no elaboration by me. No destructive behavior by "individual" social insects is unresolved. There
doubt some of those concerned with suicide prevention will fault any are at least two theoretical alternatives to the treatment of nonrepro-
attempt to consider suicide as an adaptive process. But it is clear that ductive worker members of social insect societies. One is to treat them
in the case of some human hunter-gatherer societies the kin selection as biological individuals like any other animal population and to include
hypothesis for suicide as an adaptive process has both strong logical them in comparative discussions. This tactic was freely used by E. O.
and empirical support. Wilson (1975) in his now classical book, Sociobiology. An alternative is
Human kin selection hypotheses can be placed in direct contrast to to consider such "individuals" as in every sense involuntary evolution-
Skinner's (1969) contention that suicide must be learned because it is ary slaves of the controlling reproductive caste. This alternative
always maladaptive. I single out this evolution-culture dichotomy suggests that these individuals have no control over their biological
because Skinner's alternative automatically throws support to explana- destiny other than that imposed by their genetic and facultative control
tions lying on the cultural side, an alternative that critics of genetic of the reproductive members of the colony. I strongly favor this
determinism (hard-core sociobiology) should be quick to seize upon. alternative and would treat a worker honeybee, for example, as no

more in control of its destiny than is the stinging cell of a hydra or an suicides. If it can be shown that there is any significant increase in
anemone. The genetic reproductive potential of asexual insect workers reproductive success for surviving individuals, even if it is massively
acts through the survival of the parent and since the parent controls less than that for individuals who do not express such behavior,
the progeny and its actions, it seems incorrect to include the behavioral selection against the maladaptive (suicidal) behavior will persist. The
actions of these asexual workers as comparatively analogous to social more difficult part of the equation to analyze is: what would have been
phenomena in general and suicide in particular. In other words, a the benefit to relatives if the suicide had been successful? The difficulty
worker honeybee is not committing suicide when it stings a bear on the of accurately making such a measurement leaves this analysis in the
nose, pulls away its stinger, and loses its life, any more than a lizard is area of speculation, probably forever.
doing so when its tail breaks away in the extreme emergency of an In conclusion, I must answer the question of animal suicide with a
attack. Energy is gone, but reproductive potential remains if the probable no. Few if any animals in nature commit completed suicide.
evasive tactic is successful. Suicide of a biological individual is not The closest analogues appear to be emigrations of insects and
involved. self-sacrifice of doomed insects to predators; both are poorly studied
Extreme parental investment. Parental investment in progeny is phenomena. In captivity, the expression of subsuicidal behavior is, by
made in proportion to the optimum fitness of progeny. If the optimum contrast, commonplace. Perhaps the greatest insight an animal-human
strategy is to maximize reproduction once and for all, as is the case for comparative analysis offers is that humans have, in their transition from
those salmon species that spawn and die, then this is a route nature to a form of self-imposed captivity, voluntarily created out-
life-history strategies can take. A spawning salmon can be said to be of-evolutionary-context environments where zoolike self-destructive
committing suicide. But it is an adaptive strategy and an evolutionary behavior is commonplace.
choice of alternatives. Some salmon species have evolved, as an Humans were probably a highly kin selected species. Thus, a part of
alternative, a less complete allocation of energy resources to spawn- their behavioral repertoire may have been shaped by earlier evolution-
ing, a return to the sea, and the possibility of returning in subsequent ary conditions where self-sacrifice via suicide was an adaptive
years to spawn again. All of this is potentially explicable in terms of process. But study of contemporary suicide by humans may be no
classical natural selection, and a learning process is not implicated. All more meaningful to evolution than is the study of analogous behavior
other well studied examples of parental investment (Trivers 1972) by animals in zoos and laboratories.
seem to tit well into an adaptive evolutionary framework and thus lie
outside the current discussion.
Antipredation behavior. Animals sometimes take unnecessary by L. D. Hankoff and William J. Turner
risks in the presence of predators. All mobbing behavior and alarm Department of Psychiatry and School of Medicine. State University ot New York at
calling falls into this category. An extensive literature supports adaptive Stony Brook. Stony Brook, N.Y. It 794
explanations of these phenomena involving primates, birds, and
A nontheory of suicide
mammals.
A different logic has been advanced to explain the relationship of the The heterogeneity of suicidal behavior is a commanding issue in any
risk-taking behavior of certain butterfly species to their vertebrate effort to understand the many forms that suicidal behavior takes. The
predators. Adults or larvae may offer themselves to predators when classification of suicidal behavior must take into consideration the
senescent (postreproductive) (Blest 1963) or when doomed by surely broad range of individuals, acts, and situations involved. Considering
lethal internal parasites (Shapiro 1975). One adaptive benefit of such the basic parameters of age, sex, and fatal versus nonfatal forms, a
behavior may be the indoctrination of predators in the case of matrix of eight cells based simply on a dichotomization of each of
distasteful and colorful moths (Blest 1963). Shapiro's clever explana- these three variables is possible. Obviously, a young woman who
tions of the apparently suicidal behavior of parasitized nymphalid carries out an act of attempted suicide poses a picture and a challenge
butterfly larvae seem plausible, but the phenomenon is largely un- to our understanding radically different from that of an old man who
studied, and a more secure understanding of it requires further study. commits suicide.
DeCatanzaro does not make a great distinction between self- The age distribution of suicide varies with sex in most western
mutilation in captivity and in nature. The paraphernalia (artifactual countries. Among men suicide risk increases with age, whereas among
basis) for suicide is more available in captive situations. In addition, it women the age rate curve is a flatter one which falls off in the fifties.
seems likely that many of the self-destructive propensities of captive The risk of suicide is invariably higher among men than among women.
organisms are the result of the expression of behavior patterns that In the case of attempted suicide, both of these relationships are
would be adaptive in nature. For example, wolves and other canines changed. Attempted suicides occur more frequently among women
range widely daily, regardless of their hunting success (except when than men, and the risk of an attempt decreases with age for both
denning). This insures that they do not overhunt local areas, that they sexes, by and large. The peak for suicide attempts is around 20 years
encounter territorial transgressors, and that they develop an accurate of age for both men and women, in contrast to the older age peaks for
learned inventory of the prey within their space. The same behavior, fatal suicides in both sexes.
expressed in captivity, may result in a pacing behavior resulting in Impressive efforts at a total view of suicidal behavior date back
physical damage and an early death. Other forms of subsuicidal many decades. Durkheim's (1951) volume, which was first published in
behavior by captive primates are also associated with an environment French in 1897, presents a four-way classification of suicidal fatalities
that does not include the social or physical conditions that would that has stood the test of time and remains valuable in stimulating
reorient this self-destructive activity in nature. continuing research. Durkheim divided suicide into egoistic, anomic,
Most contemporary human populations are to some extent out of altruistic, and fatalistic types, relating each to a variety of real situations
evolutionary context and live in self-imposed captivity. I therefore feel that he observed in his own data from France in the nineteenth century
that the only viable comparison with animals is to caged creatures. and a wide variety of other sources dating back to the eighteenth
DeCatanzaro notes that there are age-specific expressions of century and covering all of Europe.
behavior. Few evolutionary oriented biologists would deny that this is Durkheim, as the father of modern sociology, emphasized and
true and that for many animal species, and probably for humans, this clarified social influences in the genesis of suicide. His understanding
is, in part, a heritable set of traits. Can this lead to the expression of remains a valuable vantage point for our current view of suicide. While
maladaptive behavior at any age? By definition, maladaptive charac- the emphasis in most modern approaches has been on psychopathol-
ters cannot be selected for and, given even marginal reproductive ogy, and the refinements in diagnostic classifications in psychiatry
prospects, there is going to be selective pressure against suicide. To have helped establish firm links with suicidal behavior, the influence of
evaluate the costs and benefits of suicide it will be necessary to social factors remains quite evident in many cases.
evaluate individuals who have made unsuccessful attempts at suicide The most dramatic example of suicide in the past decade was the
(e.g. by leaping from tall buildings and surviving) and to compare this Jonestown, Guyana, incident which involved over 900 individuals.
with the positive and negative benefits to relatives of completed Clearly, this large group of men and women of all ages must have

included many who were entirely free of psychopathology and who behaviors. This is more than hair splitting over the fact that the word
were dedicated to suicide on the basis of social and peer-group suicide is a noun and not a verb. Thinking about suicide as a single
influences. Durkheim allowed for combinations of his four suicidal behavior creates an illusion of a single behavioral act, which may be
categories, and it is very likely that the Jonestown mass suicide would convenient for discussing the sociobiological transmission of the
have been classified by Durkheim as an example of altruistic-fatalistic activity but is a serious distortion of reality. In a more sophisticated and
suicide. Durkheim used the term "altruistic" in a specific sociological appropriate analysis suicide would be approached as the outcome of
sense, roughly equivalent to our understanding of socially prescribed any of a large variety of motivations, in response to varied antecedent
and formally defined behavior, not as altruism in the sense of behavior conditions, with totally different behavioral responses used to produce
devoted to the welfare of an individual or a group. that outcome. To use so vague and overinclusive a term in an attempt
Durkheim was aware of suicide in traditional and primitive societies at scientific analysis is bound to lead to great confusion - and it does.
although his data were obviously not nearly as exact as those relating The oversimplification of the concept of suicide provides a shaky
to the industrialized countries of Europe. The data on suicide in foundation to the models deCatanzaro puts forward for discussion.
underdeveloped, traditional, and primitive societies indicate that het- Our understanding of suicide is subject to further distortion by his view
erogeneity considerations hold for these societies as well. It is often that it does not have environmental outcomes for the victim. DeCatan-
assumed that primitive society, free of urban and industrial pressures, zaro states that "completed suicide is perhaps the most unambiguous
is relatively free of suicide. This is not the case, and relatively high form of self-destructive behavior, in that it is difficult to conceive of any
suicide rates have been associated with specific groups. For example, secondary gain that could accrue to any individual not surviving his
B. Malinowsky spoke of the frequent suicides among the Trobriand suicidal act." This statement ignores the human ability to experience
Islanders he studied and Peter Freuchen made the same observation "gains" from symbolic acts, fantasized events, future projects for the
of the Eskimos with whom he lived for many years. self, and so on. DeCatanzaro's view of gain is much too narrow to
Suicide and biological factors. The clear-cut findings relating encompass people's cognitive ability to go well beyond concrete
suicide to biological factors are few in number. No genetic pattern in reinforcers. Such a statement also ignores the literature on suicide.
terms of inheritance has emerged in relation to suicidal behavior. On Much of the research on suicide postmortems has demonstrated the
the other hand, the relation of suicidal risk to affective disorder is well imagined impact the suicidal act is meant to have on the world left
established. The affective-disorder patient has a risk at least fifteen behind. Biographical and novelistic presentations of the suicidal experi-
times that of the general population of dying by suicide. The few ence are even more persuasive in their support of just how very
biochemical studies of suicidal individuals suggest that there are important the "gains" are to the person contemplating or committing
biogenic amine changes in specific diencephalic and brainstem areas suicide. Clinicians who have worked with potential suicides and suicide
that may distinguish the suicidal from the nonsuicidal brain. However, attempters are quickly made aware of just how important is the
such findings are few in number and may not be distinguishing a ideation about the aftereffects of the act. It is the rare individual who
suicidal affective-disorder patient from the nonsuicidal affective- does not have a desirable outcome in mind - that is, a strong
disorder patient. reinforcement for the event. Not infrequently there is a cognitive
Any correlation between suicide and survival of the human species distortion of the fact that they themselves will no longer exist after the
does not appear strong in view of the fact that most suicidal suicidal act to enjoy the outcome. The clinician's therapeutic activity
fatalities occur among individuals who are beyond the child-bearing must often be directed toward correcting this distortion. Put in this light,
period. It seems unlikely, therefore, that suicide affects mankind's suicide is clearly more consistent with learning principles than the
survival one way or the other, in evolutionary terms. While genetic and unexplainable, simple behavioral act without gain as proposed by
biological factors may be at work, the likelihood of determining such a deCatanzaro.
connection depends upon an understanding of the heterogeneity of Even if we could accept deCatanzaro's view of suicide as behavior
suicidal behavior. Does our biological theory relate suicide to biological with no gain for the person, we would have to question strongly his
correlates in the old or the young, in men or women, or in fatal suicides defense of a genetic transmission of that behavior. The weight of the
or attempted suicides? Genetic factors may be significant in terms of empirical evidence is much more supportive of a sociopsychological
some suicides, but it is necessary to isolate that subgroup of suicidal determination of suicide. Differences in suicide rates in different
behavior in order to demonstrate significant correlations and arrive at countries, religious groups, and times of the year (seasons, holidays)
any proof of the theory. Because of the wide variety of suicidal and in response to social upheaval or stress (war, famine) all support
behaviors and suicidal individuals, any theoretical framework is environmental determination. The sociopsychological argument finds
doomed to some success since there is something in suicidal behavior strongest support in the changes in the suicide rates in cultures
for every point of view. The totality of theoretical efforts in terms of undergoing dramatic social change such as industrialization, urbaniza-
suicide emphasizes again the heterogeneity of this form of behavior tion, and political upheaval. In these examples the genetic pool is the
and the need to seek specific theories for specific and operationally same yet the rate reflects environmental determinants at both social
defined subgroups within the broad group of suicidal behaviors. and individual levels.
DeCatanzaro proposes four models for suicide, each representing a
different perspective on the act. Each perspective is well presented,
by Morton G. Harmatz with strengths and weaknesses clearly stated. While he argues in his
Department ol Psychology, University ol Massachusetts. Amherst. Maes. 01003 conclusion that these models are not mutually exclusive. deCatanzaro
does favor the fourth one as the best support for a sociobiological
The biological perspective on suicide: to be or not to position. This model indicates that suicide becomes possible when
be - is that sociobiology? individuals are less capable of reproducing their genes. This model
The examination of suicide from an evolutionary and sociobiological attempts to integrate evolutionary and environmental factors by
perspective has led deCatanzaro to some interesting speculations proposing a release from constraints against suicide when the gene
about suicide and the special challenge it poses for sociobiological pool will not be affected. It is not clear however, how this change is
theory. The models for suicide that are presented and compared are mediated in the individual, nor how such a releasing mechanism would
meant to provide a springboard to further research and analysis of the be communicated within the species.
problem. The discussion is stimulating, but there are problems at the As a central argument for the fourth model, deCatanzaro states that
very core of the analysis that raise serious questions about the value of suicide occurs in those individuals who are expecting little improvement
a sociobiological analysis of suicide as presented here. in their future. It could well be argued that suicidal ideation is actually an
One problem emerges at the very beginning of the discussion with attempt to create a desirable future from a highly undesirable present.
deCatanzaro's use of the term suicide. He approaches it simply as a This is suicide as a "rite de passage" to a better existence, unrealistic
behavior. Suicide is not a behavior in the formal sense of that term, but as it may be. A clinical sign of concern is the sudden change to a joyful
rather a descriptive label for the outcome of a large variety of other mood in the potentially suicidal person, which is frequently the time of

greatest danger. Suicide is seen as the solution, not the problem. absent, resulting in a diminution rather than an enhancement of the
Again, a learning interpretation seems to fit the data better. individual's life.
We have argued that much of the confusion in the discussion results Suicide In modern and ancient times. A second comment
from the way deCatanzaro uses the concept of suicide. These worth making is that we may be distorting our analysis of suicide if we
criticisms do not necessarily mean there is no sociobiological compo- focus solely on suicide in modern times. In deCatanzaro's discussion
nent to self-injurious behavior. An alternative approach is to move of Model IV ("Evolutionary tolerance of Suicide"), he notes that suicide
away from the confusion created by the concept of suicide and look to is much more common in the elderly. In primitive times, the life
a more general state of the organism as the mediator of these potential expectancy of humans was only twenty years or so. Few people
responses. Depression may well prove to be the more appropriate survived into what we now call adulthood. Thus, suicide was probably
mechanism for understanding suicide as a sociobiological event. much rarer in ancient times (let's say five or ten thousand years ago)
Evidence is mounting that humans and other animals appear to have a than it is today.
neurophysiological capability for experiencing depression. The depres- The advanced technology of modern times, leading to the survival of
sion appears to be activated by any of a large number of environmen- people into what we now call "old age" may have led to the
tal and physical events. This could be the biological mechanism for the appearance of pathological behaviors because of the "abnormality"
removal of weak or damaged organisms from the group and thus have of such long-term survival. It may be that evolutionary processes will
selective value for the species. The depression could lead to any of a eventually lead to the selection of genes and individuals more suited to
number of behavioral outcomes based on social and individual learning modern environments. Perhaps the existence of behaviors such as
- self-injurious behavior being but one possibility. Other organisms may suicide simply indicates that technology advances much more quickly
simply "waste away," while still others may succumb to natural than the natural selection of genes appropriate to the environment.
predators. The work on learned helplessness could easily be incorpo- Similarly, suicide may have been useful as an altruistic behavior in
rated into this model. The phenomenon of depression provides a much the sense of deCatanzaro's Model III ("Suicide as altruism") in ancient
better fit to both the arguments raised by deCatanzaro and may well times. In primitive times, food supplies were most likely often sparse,
prove to be the better construct on which to expend the research effort leading to such behaviors as suicide among the elderly and infirm, the
called for by deCatanzaro in his conclusion. murder of neonates born soon after the birth of another child, and
other behaviors that served to limit the population.
by David Lester The advances in technology that have increased food supplies for
Richard Stockton State College. Faculty of Social and Behavioral Sciences. most nations have served to make these population-limiting behaviors
Pomona. N.J. 08240 obsolete and unnecessary. However, changes in the genes may lag
behind the rapid advances of technology. This speculation about the
The categorization of suicide living conditions in ancient times has been applied by others to the
Suicide is far from easy to categorize. What kind of behavior is it? Does existence of an incest taboo in modern times (Slater 1959) and, in a
it stand by itself in a class of behaviors, or is it rather a manifestation of more analogous situation, to the benefits of having genes that predis-
a broader class of behaviors? On the whole, deCatanzaro discusses pose an individual to schizophrenia (Jonas and Jonas 1975; Osmond
suicide as a behavior standing alone, although at times he mentions and Hoffer, n.d.).
the relation of suicide to attempts at suicide and the relation of suicide
to depression. The categorization of suicide has important implications
for the biological perspective that deCatanzaro proposes. by F.V. Wenz
For example, let us assume that suicide is an aggressive behavior, Department ot Sociology. University of South Carolina. Spananburg. S.C. 29303
and therefore appropriately compared with other aggressive behaviors

Heredity, environment, and culture in suicide
such as homicide and assault, as Henry and Short (1954) have
proposed. We would then discuss the relevance of aggression for the I wish to compliment Dr. deCatanzaro for an excellent review of the
survival of the individual organism's genes, and deCatanzaro's Model II literature on the present state of biogenetically oriented research and
("Suicide as pathology") would be concerned with environments that theorizing about various manifestations of self-aggression. Since my
make externally directed aggression impossible, thereby facilitating the own work has not brought me into direct contact with this research, I
appearance of internally directed aggression (Henry and Short 1954). should like to bring out some of the weaknesses of this orientation with
Or, alternatively, let us categorize suicide as an extreme manifesta- respect to suicide. My remarks are addressed against the strict genetic
tion of depression, as deCatanzaro does in his section "Conclusions determinism of human behavior in general, and suicide in particular.
and research suggestions." We would then be concerned with the The writer states that "suicidal behavior acts to decrease rather
relevance of depression (and its underlying physiological bases) for than increase the biological fitness of the behaving organism." Well,
the survival of the organism's genes. Again, suicide may be seen as a physicians, other university graduates, and the gifted (i.e. intellectually
rare pathological outcome of such a process. superior) are regarded by many investigators as high suicide risk
Many investigators consider suicide to be a symptom of psychiatric groups. On the other hand, prisoners, hospitalized mental patients, the
illness. Estimates of the number of suicides who are psychiatrically ill military, and the never married, who comprise about five percent of the
vary widely. DeCatanzaro cites a figure of one-third, yet other investi- population 15 years of age and older, are also high suicide risk groups.
gators set this estimate as high as 94 percent (Lester 1972). To Does that mean that these diverse groups are biologically unfit? Is
consider suicide as merely a symptom of psychiatric illness would shift there a ubiquitous genetic precursor to suicide?
our approach to the models proposed. The oft-quoted statement by Linus Pauling and Ralph Gerard that
Broadening the concept of suicide would also have certain implica- "behind each twisted thought lies a twisted gene" gives succinct
tions. Viewing completed suicide as similar to attempted suicide and expression to the convictions of sociobiologists that suicide will
other minor manifestations of suicidal tendencies would make it ultimately be understood, at least in part, as reflections of basic
possible for us to place suicide in a context in which it could be seen as genetic evolutionary processes. DeCatanzaro's hypothesis that suici-
beneficial to the organism. Commonly, suicidal episodes are seen as dal individuals are those least capable of promoting their genes may be
crises (Tabachnick 1970), and crises can be seen as periods for correct. But while our ignorance still exceeds our knowledge by a wide
potential psychological growth. Too frequently, psychologists view margin, there can be no doubt that continuing biological scrutiny of
periods with negative affect as bad for people, and they try to eliminate self-injurious behavior may yield some increments in theoretical under-
them. The viewpoint of existentialists and some humanistic psycholo- standing.
gists, in contrast, is that such periods must not be avoided, but rather If we ask, "Why do newborn ladybugs eat their brothers?" or "Why
explored in depth, so that the individual can grow from this experience does the female praying mantis sometimes eat her mate?", we are
(Laing 1967; Lester 1977). In this case, suicide may be seen as a usually satisfied with the answer, "Well, that's the way they are." When
crisis and growth period, in which helpful environmental supports were we ask ourselves, "Why do human beings kill each other or themselves

Commentary / deCatanzaro: Biology of suicide
at so great a rate?" such a dispositional explanation does not satisfy disease or human behavior is always legitimate. For example, our
us. We do not wish to hear that human beings kill themselves because present knowledge of the association between blood groups and
"that's the way we are." So we search for clues that seem to trigger disease indicates that few, if any, genes are entirely neutral from the
the suicide in the genes or the environment. standpoint of evolution. They may have positive or negative selective
The possibility that animals commit suicide is one of those intriguing value for survival. In the individual case of suicide we may be able to
suggestions that has been proposed regularly by scientific observers. afford to ignore genetics, but to do so in terms of populations might be
Research has demonstrated that animals do die under the impact of a mistake.
environmental and psychological stress. Furthermore, animals may
hasten their own deaths through specific self-injurious acts. It appears
very likely that suicide may occur in animals in the absence of a
knowledge or concept of self. by Richard D. Wetzel
My general social science background has led to the tentative Department of Psychiatry, Washington University School of Medicine and the
conclusion that no unitary explanation of the nature of suicide will Jewish Hospital of St. Louis, St. Louis, Mo. 83tW
contribute significantly to the advancement of research on suicide. A

Suicide: the need for a cognitive perspective
biological view of suicide oversimplifies both social Darwinism and the
human condition by reducing suicide, however defined, to the level of Research shows that suicide is usually a premeditated act (Stengel et
genes. What one must not lose sight of is the important fact that, while al. 1973; Sainsbury 1955). Since it appears to be the result of a
genetic data may provide a basis for the understanding of the human decision, some attention has been directed to how suicidal people
species as a species, it cannot go very far in explaining attitudinal and think. Neuringer and his associates (Neuringer 1968, 1973, 1974,
behavioral differences among human populations. Humans are 1976; Neuringer and Lettieri 1971; Levenson 1974) have suggested
animals, but they need not necessarily behave like other animals. that suicidal individuals have a lifelong tendency to think dichotomously
Human evolution is unique insofar as it is regulated by an additional (in either/or terms) about all concepts and that this tendency leads to
unique variable, namely culture, rather than just the usual two variables, their suicidal decision. Wetzel (1975a, 1975b, 1976b) was able to
heredity and environment. Any explanation that attributes suicide to replicate Neuringer's findings but also showed that the tendency to
heredity only, to social factors only, or to environmental factors will be think dichotomously applied only to concepts relevant to the crisis (e.g.
unproductive. Explanations of the occurrence of suicide in individuals, life, self). He further showed that evaluations of life and self changed
and of their distribution in populations should take into account over time. These changes were highly correlated with changes in
knowledge of how heredity, culture, and environment interact to suicidal intent, depression, and hopelessness.
produce the observed distributions. DeCatanzaro shows he agrees Beck and his associates (Minkoff et al. 1974; Beck et al. 1975) have
with this notion by stating that "it is thus likely that behavioral suggested that suicide is a function of a sense of hopelessness. They
development involves complex interactions between learning and have shown that suicidal intent is more closely correlated with hope-
innate variables." However, the author overstates the case in the lessness (as they have operationalized it) than with depression. Their
statement that "the innate component should reflect countless genera- results have been replicated using different measures of depression
tions of selection for adaptiveness, and we should thus be able to
and different populations (Wetzel 1976a, 1980). Beck (1972) has
account for most behavior in terms of evolutionary contingencies that
described the cognitive triad in depression as a negative view of
affect it."
oneself, a negative view of one's current life situation, and a negative
I believe that a biological framework is necessary for an understand-
view of the future (hopelessness). While it is clear that cognitive
ing of human behavior, but it is important to investigate the role of
aspects of depression correlate more with suicidal intent than do other
culture more adequately in population and genetic studies of human
components of the depressive syndrome, no research to date shows
behavior. For instance, operational definitions of culture are needed so
that any one member of the cognitive triad correlates more with
that cultural variables may be quantified, and persons and groups can
suicidal intent than the others. The available data show that all three
be classified accordingly. The major problems we now face in research
are highly correlated with suicidal intent and load on the same factor
on suicide result from the fact that we do not have adequate informa-
(Wetzel 1976b).
tion about the necessary and sufficient conditions for its occurrence.
Beck's theory suggests that depression (and hopelessness as an
Research that would uncover a necessary condition or single factor
aspect of it) may be a function of values (schemata) learned early in life
that must be present for a specific type of suicide to occur, regardless
and maintained through cognitive distortions of experience (Beck
of whatever other factors are involved, would be a major contribution
1976; Beck et al. 1979). No data are available on this point. No family
to any of the four models discussed by deCatanzaro. Suicide occurs in
studies on cognitive distortions have been done. There is an obvious
people who have little in common - except their behavior. In fact, the
need for family studies looking at the relation among "irrational"
cause(s) of suicide are so different that we suspect that there may be
many kinds of behaviors masquerading under the single label of beliefs, cognitive distortions, and depression/hopelessness/suicide.
suicide. Since it is clear that depression runs in families, one would expect
these cognitive factors (if causally related) to be familial as well. The
Genes are responsible for our capacity to acquire culture. Genes appropriate research design and data analysis might allow one to
also form the basis for other capacities, the wide potential range of distinguish multifactorial genetic transmission from nongenetic familial
behaviors. Different human populations may have somewhat different
transmission. Currently available data exclude single-gene models for
potentials, but it must be borne in mind that while genes may provide
the transmission of unipolar affective disorder.
different probabilities for behavior they do not produce the actual
Cohort analysis of Canadian and American suicide rates shows that
behavior itself. Different environmental conditions can produce the
there are marked cohort and secular trend effects (Hellon and
same behavioral results with different genetic backgrounds. Individuals
Solomon, in press; Murphy and Wetzel, in press; Wetzel and Murphy, in
with very similar genetic backgrounds can be very different.
preparation). The data indicate clearly that cohorts show consistently
Furthermore, humans can overcome certain genetic disabilities high or low rates predictable from the cohort's rates at ages 15-19.
through culture. If there is a genetic predisposition toward suicide, this While hardly confirming Beck's theories about the time of origin of
predisposition can be modified by culture. If more and more humans cognitive structures and processes leading to depression/hopeless-
come to wear glasses as the result of "bad eye genes," the success ness/suicide, these sociological findings are obviously compatible. It
of the human species will not be threatened. Eyeglasses are a would be intriguing to have the data to assess the correlation between
legitimate substitute for better eyes. cohort and other sociological predictions of suicide rates and the
Whatever may be the final verdict on the heredity-culture-environ- cognitive variables implicated in suicide.
ment-suicide controversy, there remains no doubt that genetic studies An attempt, like that by deCatanzaro, to choose between models,
are as important as psychological and sociological studies of suicide. while fascinating, is clearly premature. The social, familial, and genetic
Any inquiry into the nature and significance of genetic variables in any bases for the cognitive content and processes of suicidal patients

Response/deCatanzaro: Biology of suicide
need to be worked out in detail before conclusions about suicide's about overall proportion of shared genes is a common technical error.
evolutionary significance are attempted. The overall frequency of the gene is irrelevant. It is merely necessary
that the gene frequency among the recipients of the altruistic behavior
be greater than among the population at large (Dawkins 1980; Wilson
by David Sloan Wilson 1980). However, kin and group selection may still be important under a
Division of Environmental Studies, University ot California, Davis, Calif. 95616 variety of circumstances. If the prospects for personal fitness are
extremely low, even a small benefit to recipients may be justified. A
Suicide, beanbag genetics, and pleiotropy society can make suicide adaptive if it threatens the disgrace of the
The idea that a given trait or behavior is coded by a single gene was whole family as the only alternative, or glorifies the family as a result
labeled "beanbag genetics" by Mayr (1963). Although long ago (Ghiselin 1974). Suicide may be justified if it carries large-scale
abandoned by geneticists, the beanbag approach is still used by many implications for societal change.
ecologists, ethologists, and sociobiologists, mostly for simplicity and Many of the above points are also made in deCatanzaro's interest-
for lack of a general alternative model. Suicide and other self-sacrificial ing target article. This note is intended to emphasize that most of his
behaviors pose a problem for the beanbag approach because they are arguments can be easily incorporated into a purely biological model,
so obviously selected against, unless they benefit others who share the when a realistic genetic structure is considered. Suicide may be
same gene. However, many of these problems disappear when more advantageous in itself or a necessary cost of behavior that is advanta-
sophisticated genetic models are employed. geous when averaged over all circumstances. In both cases, the
Fundamentally, we wish to explain how evolution can produce improbability of change emerges as a critical variable, either by
maladaptive individuals, or maladaptive traits in otherwise well adapted predisposing the individual toward the benefit of others, or by causing
individuals. The three most obvious candidates are novel environ- suicidal situations to have very little "weight" in the calculation of
ments, genetic load, and pleiotropy. That individuals may behave average fitness. I suggest that suicide can best be understood, not as
against their own interest in environments to which they are not an isolated behavior, but as one manifestation of humans' complex
adapted is straightforward and needs no more discussion. That sexual adaptive machinery.
reproduction can produce bad genetic combinations also needs little
elaboration. The concept of pleiotropy recognizes that a gene can
code for not one, but several behaviors. It could conceivably explain
suicidal tendencies in "normal" individuals when they are placed in
certain unusual circumstances, and is therefore more interesting. Author's Response
Simply put, many adaptations involve both costs and benefits, and
natural selection is expected to maximize the net gain. Maladaptive by Denys deCatanzaro
Department of Psychology, McMaster University, Hamilton, Ontario, LBS 4K1,
behavior may be selected if it is a necessary part of a larger behavioral
Canada
complex whose overall effect on fitness is positive. As one pertinent
example, it is possible that an unavoidable physiological trade-off
exists between vigor and longevity. Under some conditions a short- Human suicide: toward a diathesis-stress hypothesis
lived, vigorous animal will leave more offspring than a long-lived, At the outset, I would like to restate my original intentions in
sluggish animal, in which case early senescence is the price paid for an outlining alternative biological interpretations of human
exuberant youth (e.g. Stearns 1976). suicide. My analysis was not intended to provide definitive
A parallel argument for suicide would be as follows. Consider a answers with respect to the biological anomaly of suicide -
species whose individuals find themselves placed permanently in one that would surely be premature. Rather, my purpose was to
of two "circumstances" with probabilities p, and p2 (= 1 — pt). The stimulate examination of this phenomenon at a level of
maximum fitness that can be obtained within each circumstance is F, analysis radically different from that of the existing literature.
and F2. The individuals possess a single behavior pattern that allows I am pleased that my target article has succeeded in eliciting
them to realize a fraction q, of the fitness in circumstance 1, and q2 of some insightful elaborations and criticisms, but I must also
the fitness in circumstance 2. The expected fitness of the individual is draw attention to some unfortunate misinterpretations.
therefore The level of analysis that I undertook has been notably
W - q,p,F, + q2p2F2.
absent from the sociological and psychological literature on
suicide. There is a need for a framework that integrates the
As an example, let q, - 1 and q2 = 0. In other words, individuals are otherwise apparently disparate findings in these areas; such a
maximally fit in circumstance 1 and suicidal in circumstance 2. Several framework would surely require a global perspective rooted
interesting conclusions emerge from this simple model. First, suicidal in, or at the very least concordant with, biological science.
behavior will have only minor effects on fitness if either p2 or F2 is very The sociological literature indicates remarkable consistencies
small; that is, if either the probability of encountering, or the fitness in the social ecology of suicides, yet there has been no
associated with the second circumstance is minute. Of course as a complete attempt to account for the genesis of this social
"beanbag" trait, q2 should increase no matter how small its effect on ecology. In psychology, behavior has long been viewed in
fitness, but as a pleiotropic trait q2 will not increase if q, decreases terms of its orientation toward reinforcers and away from
along with it. (In ecological jargon, this model is identical to the punishers, which implicitly, but not explicitly, addresses the
question of whether a species should specialize in a coarse-grained relationship of behavior to biological fitness. Both disciplines
environment; Levins 1968). appear to be in need of infusions of some of the fertile notions
Second, redefine p, and p2 to be the proportion of time spent by all that are being developed in the areas of sociobiology and
individuals in the two circumstances. Now q2 - 0 is highly disadvanta- behavioral ecology.
geous, no matter how small the value of p2 and F2, because suicide For the sociobiologist, human suicide may present a special
prevents a return to the more desirable circumstance. To summarize, challenge. No other behavior appears to be so contrary to the
this simple model predicts that suicide will be most prevalent in principles of natural selection, at least when analyzed at a
circumstances that are infrequent and contain few prospects for superficial level. Such problematic phenomena may provide
reproductive success and little opportunity for change. This of course tests of the limits of sociobiological reasoning. Indeed, it is
matches the observations, indicating that suicide does not pose a undoubtedly wrong to attribute adaptiveness to all the fine
problem for evolutionary models of human behavior, although nonevo- details of behavior; there is no exact correspondence between
lutionary models could also be constructed to explain the same selective pressures and behavior. Our task is to find out
patterns. through research and analysis how precisely human behavior
With respect to kin and group selection, deCatanzaro's statement is determined by natural selection.

My outline here follows that of the target article, and I diagnostic categories is tenuous. Such nosological systems
address each commentator's suggestions as they arise in this have almost always failed to bring order to the diversity of
sequence. At times this sequence may break down as I discuss syndromes of abnormal behavior. In considering schizophre-
relationships among the models. At the end, I present a nia, depression, aphasia, child psychosis, suicide, or almost
tentative synthesis that integrates the diverse notions any of the major forms of behavior disorder, we must not
presented by myself and the commentators. forget that each individual is different. Different etiologies
may yield convergent symptomatologies; more often, disor-
Evolution and behavior. My suggestion that innate ders result from a synergism of diverse factors. I doubt that
influences on behavior should reflect generations of selection Durkheim's subclassifications of suicide, being founded on a
for adaptiveness and that we should thus be able to account small and old data base, retain much utility or construct
for most behavior in terms of evolutionary contingencies is validity. The diversity of suicidal behavior is probably best
questioned by Wenz and Duke. I was, however, allowing for accounted for by differential concentrations of many dispa-
the influence of learning here, as should be clear from the rate determinants acting convergently or synergistically in
context of this statement. Also, I was simplifying to set up the different individuals. We must accordingly consider a broad
problem that suicide presents biologically. The subsequent range of etiological factors and their interactions, a unique
discussions, especially those in the pathology model, define combination for each individual.
limits on this statement. But, where the numerous patho-
genic events do not occur, this statement should hold in full. It Model I. Some novel expansions of the notion that complex
is supported by a growing literature on the behavior of and unusual cognitive events may underlie self-damaging
organisms living in the conditions in which - they have behavior are presented by Frieden. There can be no doubt
evolved. I would thus point out to Duke that there is empiri- that suicide derives in part from human cognitive and behav-
cal support for this contention in the large literature on ioral flexibility. The involvement of cognitive variables is not
behavioral ecology. inconsistent with any of the models, and it is not my intention
Harmatz has questioned my treatment of suicide as behav- to play down the roles of cognition and learning as proximate
ior. There is no doubt that what I refer to is a class of determinants of suicide. I deliberately did not provide an
behavioral patterns, with the constituents being linked by a extensive discussion of these issues because so many other
common consequence and some sort of common antecedent such discussions are already available, limiting myself instead
motivational condition (the expressed "desire to die"). In this to a brief overview and an interpretation of the biological
sense any deliberate self-injurious act appears to be just the meaning of such factors.
same as any other goal-directed behavior. Also, I did not Wetzel has argued for a cognitive perspective on suicide.
contend that the behavioral topography itself or the mode of Again, cognition is clearly involved at a level of analysis that
suicide was anything other than learned behavior; what is differs from my own; but even when we have sorted out the
interesting is the origin of the antecedent motivational condi- role of cognition in suicide we are left with a fundamental
tions and why such learning can occur. problem from a biological standpoint; this is the anomaly
Although his title is quite clever, Harmatz has missed the outlined at the beginning of the target article. Developments
point in his assertion that I have ignored a human ability to in the study of cognitive aspects of suicide will surely clarify
experience gains from symbolic acts. Such symbolic gains some of the biological issues. Otherwise, however, Wetzel s
cannot be directly translated into gains in biological fitness. comments do not really address the issues I have raised.
My concern was with biological fitness as defined by success Hamilton's discussion of suicide by nonhuman animals is a
in gene propagation. There is no doubt in my mind that very helpful examination of a seldom treated subject. His
humans are influenced by symbols; such symbols, where they conclusions lend support to the notion that attributes specific
bear on suicide, are subsumed by factors in models I and II. to contemporary human culture are responsible for suicides.
But they are not directly representative of biological fitness, This topic has implications not only for model I (where it is
although they may relate in many interesting ways to biologi- placed only for convenience) but for each of the other models
cally meaningful gains and losses. as well. As will be elaborated below in reference to other
Frieden also briefly fell into this trap in asserting that relief commentaries, contemporary human culture supports the
from pain in dying, as negative reinforcement, constitutes a existence of genotypes in ecologically novel situations. This
gain for the individual. What matters in biological terms is results in a gene expression different from the one that was
not the individual's conception of gains or losses (although this responsible for the gene's diffusion. This converges with
is of great interest on other grounds) but rather the true Hamilton's suggestion that self-destructiveness is most
impact of his actions upon gene frequency. A cognition that common in nonhuman animals when they are in captivity, as
relief from pain justifies suicide may relate to any of the well as with the notions advanced by Dawkins, Lester, and
models (perhaps especially model I), depending on the nature Wilson (and possibly Anisman and Blanchard). This brings
of the pain. us to a discussion of the pathology model (II).
The incidence of suicide. My assertion that suicide is the Model II. The involvement of stress-induced central amine
ninth most common cause of death was quite correctly dysfunctioning in suicide is suggested by Anisman. These
questioned by Dawkins, who suggested that this depends on elaborations are informative and readily accommodated or
how finely such causes are classified. This does not diminish subsumed by the pathology model. It is very reasonable that
the problem the frequency of suicide presents, however. stress-induced abnormalities in neuroregulators and neuro-
Douglas raises the important point that statistical biases are modulators facilitate or induce maladaptive behavior.
present in official statistics on suicide. This does render However, in considering this we must also consider interac-
tentative a number of the arguments based on statistical tions with cognitive variables, and how these variables in turn
tendencies in suicide, although he notes that removal of such are influenced by selective pressures and genetics. The "de-
biases in many cases supports the conclusions I have drawn. pression " construct covers a rather amorphous and complex
Hankoff & Turner have contended that suicide is hetero- set of phenomena, only some of which may relate to suicide.
geneous in nature. I have no quarrel with this notion and As Beck, Kovacs, and Weissman (1975) pointed out, "depres-
would suggest that this heterogeneity is largely responsible for sion " is not so highly correlated with suicide as is "hopeless-
the fact that the biological causes of suicide are not obvious. ness."
But the idea that a simple nosology would neatly isolate Blanchard also offers some helpful elaborations of model

Response / deCatanzaro: Biology of suicide
II, along much the same line as Anisman. The idea that a one that was very occasionally maladaptive, allowing a gene
significant number of suicides derive from endogenous supporting both to flourish. However, such a gene might
depression not primarily related to environmental stress is an ultimately fail in competition with one allowing the advan-
important consideration. It suggests a convergence of stress- tage without the cost. For pleiotropy to be involved, as Wilson
induced neurochemical pathology and familial neurochemi- states, "maladaptive behavior [must be some] necessary part
cal pathology, which may adequately characterize some of of a larger behavioral complex whose overall effect is posi-
the variance in human depression. The question it raises is tive." Thus, we would have to find linked to suicide some
how pathological conditions conducive to suicide might arise larger trait whose overall effect is positive. This subsumes my
in the absence of stress. As Blanchard mentioned, genetic point in model I, that suicide may derive from a behavioral
variance seems to be involved, although stresses from con- flexibility that is generally adaptive but that occasionally
genital factors or environmental toxicology remain possibili- engenders maladaptiveness such as suicide. In another sense,
ties in some cases. The genetic component would be accom- pleiotropy belongs with genetic load and gene recombination
modated by my statements about genetic recombination. in model II, as means whereby genes predisposing toward
Lester also offers some strong arguments in favor of model maladaptive traits can continue to be present in the gene
II. The suggestion that an increase in human longevity may pool.
contribute to suicide is quite useful. Indeed, this strengthens An example may illustrate some relationships between
the notion that gene expression in historically novel environ- pathological conditions and associative processes. Modeling
ments is responsible. The well-documented increase in suicide (Bandura 1971) is a process whereby individuals learn vicar-
rates as a function of age is concordant with this idea. As age iously by viewing the behavior of others and its consequences.
increases, the probability that a gene is expressed in novel In normal conditions, such a process is highly adaptive
circumstances increases; consequently, the probability that it because it allows the learning of complex behavioral
will predispose toward a maladaptive phenotype increases. sequences without an extended trial and error process. First,
This converges nicely with some of the ideas of Dawkins, modeling may sometimes facilitate suicide, thus permitting
Hamilton, and Wilson, which I treat below. an occasional negative by-product because the overall process
Wenz's concern about suicide in the gifted can be handled is so adaptive. Thus, due to the rareness of such a by-product,
by the factors in model II. Some constitutional factors condu- selective pressures may not have induced a sufficiently fine
cive to suicide may be concentrated differentially in different degree of adaptation to eliminate it. However, processes such
individuals because of gene recombination; this could occur as modeling may currently produce more of a negative
irrespective of the viability of other genes the individuals by-product than they did during our evolutionary history.
carry. The same stress or aberrant environment might thus Since modern culture may expose more individuals to ecolog-
affect individuals differentially. It could also be argued that ically novel environments, there may be more opportunity for
in our culture gifted individuals are often streamed into maladaptive behavior such as suicide to arise from such
demanding and competitive environments. processes.
I would like to add an illustration of how gene expression in
novel environments may facilitate suicide. As I mentioned in Model III. According to Hamilton kin selection is respon-
the target article, technology has produced many new meth- sible for most self-destructive behavior in nonhuman animals
ods for committing suicide. These methods may increase the living in their natural ecological circumstances (i.e. where
likelihood of suicide because they are readily available for their genes evolved). This might suggest a similar involve-
impulsive use. Also, many methods may overcome natural ment of altruism in suicides of humans living in more natural
mechanisms that might deter suicide; for example, barbitu- environments. This accords with Lester's attribution to
rate overdose may overcome pain mechanisms that might altruism of suicide in ancient and less developed cultures. As I
help inhibit other methods. The availability of such methods mentioned in the target article, in modern societies altruism
is historically novel and thus not something for which we appears (on the surface) to be involved in only a minority of
would be biologically adapted. cases.
This potential influence of factors favoring altruism can be
Models I and II considered jointly. That models I and II extended by considering potential interactions with factors
need not be distinguished from one another is argued by developed in model II. Some capacity toward altruism
Dawkins. Human flexibility and some of our current novel presumably exists in the human population because of factors
environmental circumstances are of course inextricably such as kin selection, group selection, and reciprocal altruism.
linked, and the differentiation was in part for convenience of Given that we now so frequently find ourselves in situations
treatment. However, model I involves the notion that human for which we are not biologically prepared, tendencies
flexibility, while highly adaptive, has as a by-product occa- toward altruism might occasionally be expressed where they
sional maladaptive, learned behavior. The major notion do not truly serve inclusive fitness. Hence, traits attributable
underlying model II is that individuals now frequently to factors of model III may be expressed in abnormal condi-
encounter conditions for which they are not biologically tions because of factors in model II. (Could this relate to the
prepared. Quite probably, maladaptive by-products of flexi- fact that individuals committing suicide have so frequently
bility are largely responsible for unbiological environments, been isolated from kin?) This also seems consistent with the
but neither is necessary or sufficient for the other, and either last two paragraphs of Lester's commentary.
could occur independently. Model II also includes some Both Dawkins and Wilson criticized my point about
features that could have been treated separately, for example, coefficients of relationship. I realize that kin selection should
that genetic load or gene recombination might be involved in work approximately as Hamilton (1964) explained when it is
suicide. applied to the diffusion of a novel gene for altruism. I was also
Wilson presents some interesting elaborations of points not trying to repeat Washburn's (1978) point, nor did I argue
made in several of the models; these may be especially that altruism should depend on the totality of gene sharing.
relevant here. I might first state clearly that I did not suggest Undoubtedly, an individual shares more genes with close kin
that a single gene could underlie suicide. It is more likely that than with individuals from the general population. But the
suicide would be influenced by multiple genes, albeit in Hamilton-Wright coefficients provide a ratio scale when only
indirect and subtle manners. Such genes could be pleiotropic, an ordinal scale is justified. We cannot define a zero point. I
each affecting a range of phenotypes. Pleiotropy could mean doubt, consequently, that we can insert the Hamilton-Wright
that one highly advantageous trait might outweigh another coefficients, given their ordinal scale properties, into some

simple calculus to compute the tendency toward altruism. subpopulation-serving behavior, and so on. But if the pros-
Considering genes shared by kind as well as by common pects for personal fitness became very low, would not even a
descent (a distinction that becomes hazy in considering the small benefit to kin outweigh a personal loss? Similarly, if the
totality of generations), there should be a progression away prospects for both personal fitness and behavior that helped
from gene sharing between individuals who are far removed kin became low, might not group selective pressures begin to
from one another spatially, temporally, ecologically, racially, have some influence over the fate of the individual's genes?
phylogenetically, and so on. There are discontinuities in this This is a major dimension that is missing in discussions
progression, one of which may lie between kin and other belittling group selection.
subpopulation members, with the extent depending on the As Wade (1978) has cogently argued, group selection may
composition and size of the subpopulation; other discontinui- be quite powerful when it operates in the same direction
ties may lie between subpopulations, still others between as - or is simply not opposed to - pressures of individual
species, and so on. Another consideration is that if an individ- selection. I note that Wilson's comments also support this
ual or even a family is removed from a subpopulation, all of point when he says that "kin and group selection may still be
its genes, with the exception of any specific mutations, will important under a variety of circumstances. If the prospects
probably remain in the total subpopulation, albeit in smaller for . . . fitness are extremely low, even a small benefit to
proportion. The necessary and sufficient conditions for recipients may be justified." These notions again raise the
altruism toward another individual or toward the totality of question of how precisely selective pressures operate in deter-
some group would depend upon numerous parameters; this mining human behavior. A fruitful empirical enquiry might
issue must not be closed prematurely. focus on whether altruism is increased in times of personal
Let us ignore for a moment the cost of altruism to the desperation.
behaving individual. I agree with Wilson's comment that for Another important consideration is the role of human
kin and group selection to operate, it is merely necessary that culture in this process. Societies may frequently compel
gene commonality with recipients of altruism be greater than altruism from some members by placing various behavioral
with some population at large. But even allowing for the constraints on the probabilities for successful gene propaga-
discontinuities mentioned above, I am not convinced by tion. This would be true in wartime, for example, if the only
Dawkins (1976) that kin and group selection are not just two alternative to military service were severe personal disgrace.
different points along a continuum. The potential complexi- It may also characterize some of the forms of institutional
ties of gene sharing within different levels of aggregation of suicide discussed by Farberow (1972).
human individuals, especially within smaller aggregations Dawkins's suggestion that model IV be integrated with the
that characterized earlier periods of our evolutionary history, Medawar-Williams hypothesis (Medawar 1957; Williams
place estimates of "r " based on one or two generations in 1957) is most helpful. I agree wholeheartedly that this logic
some perspective. Such estimates assume random panmictic enriches model IV; I also think it can be viewed as an
breeding in a population of infinite size; such assumptions do extension of model II. Suppose that through some course of
not entirely hold for realistic human populations. events, a gene-environment combination arose that allowed
Dawkins appears to have misinterpreted my point here; the occurrence of suicide (indeed, this much has surely
the two views are not as far apart as he assumes. I am simply happened). If a gene in this combination expressed itself
less inclined to dismiss group selection. There is good reason when an individual retained a capacity to promote the
to assert that individual selection is generally more powerful continued existence of his genotype, the gene would be
than kin selection, which in turn is more powerful than deleterious and would remain infrequent in the population. It
selection among larger groups. The strongest arguments would be rapidly eliminated if expressed when the individual
against the importance of group selection rest upon mathe- was otherwise likely to reproduce and foster children.
matical models pitting group selection against individual However, if it were expressed only after reproductive age or
selection in the determination of some trait. But again the in younger individuals who were under stress and unlikely to
assumptions may be too limited. We must consider what will reproduce, and only when circumstances also prevented
occur if individual selection is rendered impotent by poor activity that indirectly promoted the individual's genotype, it
prospects for personal fitness. This brings me to the next would remain in the population. It is thus conceivable that
model (perhaps "factor" would have been a better term), genes permitting suicide, expressed only in conditions where
which elaborates my point. the prospects for personal fitness are poor, could have accu-
mulated in the population. Like the factors in model II, this
Model IV. The last model presents the notion that as describes conditions under which the individual's genes need
productive and reproductive capacity declines, there is not predispose toward adaptiveness.
progressively less evolutionary reason why the individual A caveat is necessary here. There is no claim here that
should behave in a self-preserving manner. This is really just genes directly predispose toward suicide. Such a contention
an extension of model II, because it describes an additional would be implausible given the importance of conditioning in
condition under which an individual's genes need not be human development and the consequent complexity of rela-
expressed in adaptive or self-serving ways. This condition tionships between genotypes and behavioral phenotypes in
would affect many of the same individuals as those affected humans. Rather, it suggests that multiple genes could subtly
by the factors of model II, providing further reasons why such contribute to the motivational aspects of suicide. In all likeli-
individuals might not be expected to behave in a self- hood, such genes would not be specific to suicidal behavior. A
preserving manner. weak genetic or innate basis for suicidal motivation is
This unification of models II and IV is also supported by a conceivable because of the nature of the social-ecological
number of the commentaries as I discuss below; but first I situation in which suicide appears to occur.
would like to consider the implications of this dimension for Harmatz argues that evidence that environmental factors
model III, to demonstrate further how all four models could are associated with suicide is contrary to a genetic involve-
be unified. ment in suicide. Again, he has missed the point. Gene
The importance of kin and group selection depends on the expression varies with the environmental context, as I and a
extent to which these have to compete with individual selec- number of the commentators have discussed. Thus, there is by
tion. Self-serving behavior would generally propagate the no means a conflict between the weak sort of genetic involve-
individual's genes more readily than would kin-serving ment I speak of and the evidence he cites.
behavior, which might in turn propagate his genes more than Anisman's argument that the high-risk populations of

model IV are stressed individuals points out the linkages with in model IV with the fact that offspring of forebears who
model II. The conditions necessary for the factors in model II have committed suicide exist, at least, provided that at the
appear to converge at many points with those necessary for time of suicide the forebears were not capable of promoting
the factors in model IV. Of course, the "stress" construct their genes through behavior increasing the reproductive
requires a concrete definition when used in this context. capacity of their offspring. The model hinges upon whether
Model IV really describes further conditions under which or not suicide occurs when the individual has no residual
gene expression need not be adaptive; thus, depending on our reproductive and productive capacity.
definition of "stress" and "pathology," we may ascribe such
attributes to individuals without the capacity to propagate Testing the hypotheses. According to Carr a tautology is
their genes. Also, a loss in the capacity to propagate one's inherent in my suggestion that we examine the social ecology
genes is probably associated with a substantial demand on the of suicide to determine further the relation of the act to gene
organism's coping apparatus. propagation. This is not really accurate. It is commonplace in
Duke's assertions about the logic of this section address a science to develop a hypothesis on the basis of preliminary
similar concern that the data reviewed also support a pathol- data, then to probe the same phenomena further to determine
ogy model. He suggests that we can interpret the correlations the validity of the hypothesis. Implicit in my statements is a
of stress, social isolation, and poor productivity and reproduc- suggestion that new data be collected and analyzed in light of
tive capacity with suicide in terms of a number of different the hypotheses.
causal relationships. What he does not realize is that, for One further important point can be made regarding behav-
factors discussed in model IV to hold, no particular proxi- ioral-genetic studies. Such studies only address sources of
mate causal relationship among these variables is necessary. variability among individuals and do not really test whether
All that would be required is that suicide consistently be innate factors contribute to suicide. For example, if genes
associated with behavioral dysfunctions preventing gene indirectly permitting suicide in some environments were
propagation. That association, no matter what the complexi- widely and fairly uniformly distributed in the population,
ties of proximate causation, would mean that suicide would there could be no observed difference between monozygotic
have a minimal impact upon the gene pool. I really made no and dizygotic twins. Variance among individuals is not
strong assertions about the order of causation here, nor would entirely reflective of innate (if opposed only to learned)
I need to. For example, I did not suggest that divorce causes influences on behavior, since environmental influence on
suicide, and his quote "social isolation can precipitate gene expression may be as important as associative (learning)
suicide" is taken out of context. He has missed the point here processes. As a further example, we call some fixed action
and has overlooked a number of qualifiers. patterns in nonhuman species innate because they are not
Duke's suggestion that sex differences in suicide are incon- influenced substantially by learning, yet they may vary little
sistent with model IV overlooks some relevant aspects of across different members of the species.
male-female differences in gene-propagation strategy. As is
true in many other species, human females would seem to A synthesis. My discussion is quite accurately character-
have been more prone than males to engage in kin-favoring ized by Carr as a diathesis-stress model. The majority of the
and nurturant behavior over human evolutionary history. commentators' remarks add support to such a model, as do
Consequently, a postmenopausal female (although we might the integrations of the diverse factors I have discussed in this
wonder why menopause can be correlated with severe response. I think that without prejudging the relative contri-
depression) may retain a capacity to promote her genes as butions of the various factors, we are now in a position to
long as she continues to act in some kin-favoring manner. present a relatively unitary framework that nonetheless
Duke's worries about referencing here do not undermine allows for differential causation in different individuals.
my statements. First, I did not imply that there is a decisively Let us first consider the diathesis. At one level, genes have
demonstrated relation between reserpine and suicide; I only to be involved even in suicide; all behavior is a function of
said that such a relation has been reported. Nor did I make gene-environment interactions. Inevitably, there will be some
any conclusive statements about Kallman's (1949) data. who will reduce my arguments to some gross oversimplifica-
Again, I feel that Duke has not read carefully enough to see tion, seeing absurdity in a notion that genes influence suicide.
the qualifiers. I do not see the degree of difference that he Surely, any involvement of innate factors in behavior so
implies between my statements on these two issues, taken in complexly determined as suicide is weak, subtle, and indirect.
context, and the sources he cites. Also, I believe that some No one would contend that there is a gene directly predispos-
license for secondary referencing (i.e. citing thorough reviews ing toward suicide. However, many genes undoubtedly
that elaborate, support, or qualify points) must be allowed in predispose rather directly against suicide, as must those
any endeavor operating at this interdisciplinary and global controlling physiological pain, reinforcement, and punish-
level. ment mechanisms. The largest part of the question is why
Farber's discussion seems to lend support to the notions factors such as these occasionally fail in this respect.
developed in model IV. He provides a second potential A more fruitful controversy concerns the extent to which
answer to Wenz's query about suicide in the gifted when he suicide derives from associative processes (learning) as
says that "the infrequent loss of high-quality genes would not opposed to nonassociative processes more closely related to
seem to outweigh the probably more frequent elimination" of environmentally controlled gene expression. However, this
low-quality genes. He interprets suicide as an agent of natural distinction may also be unclear, since associative processes
selection rather than as an anomalous behavior that is interact in complex ways with nonassociative motivational
contrary to natural selection. This may reduce to questions variables. This brings us to the relations between the factors in
about the locus of determination of suicidal behavior and the models I and II. The most likely relationship is that human
locus of action of selective pressures. To the extent that some flexibility and culture have, over our evolutionary history,
residual reproductive capacity is lost when suicide occurs, his given rise to our current situation, wherein human genotypes
interpretation may be correct. In line with the notion in are frequently exposed to novel environments (although some
model II that modern suicide may be increased by the other relationships have been discussed above).
exposure of genes to novel environments, suicide may indeed I have outlined several ways in which indirect innate
be one force serving to return us to genetic equilibrium with influences on suicide might occur; these have been expanded
our environment. upon by some of the commentators. A good argument can be
Regarding Farber's further comments, there is no problem made that some tendency toward altruistic self-sacrificing

References / deCatanzaro: Biology of suicide
behavior may occur in the human species. Constitutional Anisman, H. (1978) Neurochemieal changes elicited by stress: behavioral cor-
factors facilitating maladaptive self-destructive behavior may relates. In Psychopharmacoiogy of aversioely motivated behavior, ed. H.
Anisman and G. Bignami. New York: Plenum Press. [HA]
also be carried in the human population through genetic load
Anisman, H.; Pizzino, A.; and Sklar, L. S. (1980, in press) Coping with stress,
or pleiotropy. Also, when an individual is exposed to chronic, norepinephrine depletion and escape performance. Brain Re-
intractable stress, there is no basis for adaptive genetic expres- search. [HA]
sion. In this sense, stress might be equated with exposure to Bachman, J. (1972) Self-injurious behavior: a behavioral analysis. Journal of
historically or ecologically novel circumstances, conditions Abnormal Psychology 80:211-24. [DdC]
other than those in which the organism's genes evolved. There Baechler, Jean. (1979) Suicides. Trans. Barry Cooper. New York: Basic Books.
seems to be a consensus among the more biologically oriented Originally published in 1975 as Les suicides. Paris: Caemann-
commentators that the weighting of this factor should be Levy. [JDD]
increased, and I agree. The factors in model IV simply extend Bandura, A. (1971) Psychological modeling. Chicago: Aldine-Ather-
ton. [DdC]
this notion by defining additional, often similar, circum-
Barash, D. P. (1977) Sociobiology and behavior. New York: Elsevier. [DdC]
stances in which gene expression need not be adaptive. When Barchas, J. D.; Akil, H.; Elliott, C. R.; Holman, R. B.; and Watson, S. J. (1978)
an individual's capacity to propagate his genes declines, there Behavioral neurochemistry: neuroregulators and behavioral states. Science
is no ecological reason why his genotype should predispose 200:964-73. [DdC]
him toward self-preserving behavior. This last factor may in Batchelor, I. R. C , and Napier, M. B. (1953) Broken homes among attempted
turn facilitate altruism. suicides. British Journal of Delinquency 4:99. [DdC]
There can be no doubt that modern human culture Beck, A. T. (1967) Depression: clinical, experimental and theoretical aspects.
frequently places stresses on many individuals by placing New York: Harper and Row. [DdC]
them in situations for which they are not genetically (1972) Depression: cause and treatment. Philadelphia: University of Penn-
sylvania Press. [RDW]
prepared. Biological evolution has not caught up to cultural
(1976) Cognitive therapy and the emotional disorders. New York: Interna-
evolution. This is suggested by Dawkins's moth-flame analo- tional Universities Press. [RDW]
gy. Hamilton's review of self-destructiveness in nonhuman Beck, A. T.; Kovacs, M.; and Weissman, A. (1975) Hopelessness and suicidal
animals underlines how high rates of suicide may be specific behavior: an overview. Journal of the American Medical Association
to modern human culture or to animals in captivity. This 234:1146-49. [DdC.RDW]
notion is also suggested by the high rates of suicide in Beck, A. T ; Rush, A. J.; Shaw, B. F.; and Emery, G. (1979) Cognitive therapy
developed as opposed to nondeveloped countries and by the of depression. New York: The Guilford Press. [RDW]
involvement of modern technologies that facilitate impulsive Blest, A. D. (1963) Longevity, palatability and natural selection in five species
suicidal acts. Indeed, modern society has increased longevity, of New World Saturniid moth. Nature 197:1183-86. [WJH]
Bogard, H. M. (1971) Collected thoughts of a suicidologist. In Identifying sui-
as pointed out by Lester, which again exposes genes to
cide potential, ed. D. B. Anderson and L. J. McClean. New York: Behav-
conditions differing markedly from those in which they ioral Publications. [DdC]
evolved. Such stress-induced maladaptive behavior may be Breed. W. (1966) Suicide, migration, and race. Journal of Social Issues 22:30-
mediated in part by the physiological variables emphasized 43. [DdC]
by Anisman and Blanchard, and by aberrant cognitions and (1972) Five components of a basic suicide syndrome. Life-Threatening
learning experience as emphasized by some of the other Behavior 2:3-18. [DdC]
commentators. Carlson, N. R. (1977) Physiology of behavior. Boston: Allyn and Ba-
A number of commentaries have alluded to variance con. [DdC]
among individuals in terms of the characterological condi- Carr, E. G. (1977) The motivation of self-injurious behavior: a review of some
tions conducive to suicide. This variance may derive from hypotheses. Psychological Bulletin 84:800-816. [DdC]
Caspari, E. (1967) Gene action as applied to behavior. In Behavior - genetic
differential concentrations of predisposing factors and of
analysis, ed. J. Hirsch. New York: McGraw-Hill. [DdC]
aberrant or extreme environments engendering self-destruc-
Chagnon, N. A. (1979) Mate competition, favoring close kin, and village fis-
tive behavior. Undoubtedly, every individual case involves a sioning among the Yanomamo Indians. In Evolutionary biology and hu-
different combination of some subset of the diverse factors we man social behavior: an anthropological perspective, ed. N. A. Chagnon
have considered. and W. Irons. North Scituate, Mass.: Duxbury Press. [DdC]
There may be some residual occurrences of suicide that Clough, G. C. (1965) Lemmings and population problems. American Scientist
may not be "pathological, ' or that may not fit a stress- 53:199-212. [DdC,WJH]
diathesis model, but that may nonetheless derive from some Conner, R. L. (1972) Hormones, biogenic amines, and aggression. In Hor-
of the factors discussed above. These might include suicides mones and behavior, ed. S. Levine. New York: Academic Press. [DdC)
that are altruistic in nature and that truly serve the individu- Daly, M., and Wilson, M. (1978) Sex, evolution, and behavior. North Scituate,
Mass.: Duxbury Press. [DdC]
al's inclusive fitness. They may also include suicides in socie-
Dawkins, R. (1976) The selfish gene. New York: Oxford University
ties in equilibrium with their environment, although here Press. [DdC]
some "pathology " may also occur. (1979) Twelve misunderstandings of kin selection. Zeitschrift fur Tierpsy-
None of this is meant to supersede careful empirical chologie 51:184-200. [RD, DSW]
research, and at best any conclusions at this point must be deCatanzaro, D. (1978) Self-injurious behavior: a biological analysis. Motiva-
tentative. However, it would seem important that we attempt tion and Emotion 2:45-65. [DdC]
periodically to integrate at a global level relevant findings (1979) A reinterpretation of self-injurious behavior: a synergism of learning,
from a broad range of disciplines. physiological, and sociobiological factors. Manuscript submitted for publi-
cation. [DdC]
deCatanzaro, D., and Baldwin, G. (1978) Effective treatment of self-injurious
behavior through a forced arm exercise. American Journal of Mental De-
ficiency 82:433-39. [DdC]
deCatanzaro, D., and Gorzalka, B. B. (1979) Stress and male sexual behavior.
References Manuscript submitted for publication. [DdC)
Demographic Yearbook-1976. (1977) New York: United Nations. [DdC]
Akiskal, H. S., and McKinney, W. T. (1973) Depressive disorders: toward a deVos, G. A. (1968) Suicide in cross-cultural perspective. In Suicidal behaviors,
unified hypothesis. Science 183:20-29. [HA] ed. H. L. P. Resnik. Boston: Little Brown and Co. [DdC]
(1975) Overview of recent research in depression. Integration of ten concep- Diggory, J. C. (1976) United States suicide rates, 1933-1968: an analysis of
tual models into a comprehensive clinical frame. Archives of General Psy- some trends. In Suiddology: contemporary developments, ed. E. S.
chiatry 32:285-305. [HA] Shneidman. New York: Grune & Stratton. [DdC]
Alexander, R. D. (1979) Evolution and culture. In Evolutionary biology and Dorpat, T. I.; Anderson, W. F.; and Ripley, H. S. (1968) The relationship of
human social behavior: an anthropological perspective, ed. N. A. Chag- physical illness to suicide. In Suicidal behaviors, ed. H. L. P. Resnik. Bos-
non and W. Irons. North Scituate, Mass.: Duxbury Press. [DdC] ton: Little Brown and Co. [DdC]

Douglas, Jack D., The social meanings of suicide. Princeton, N. J.: Princeton Kalela, O. (1962) On the fluctuations of the numbers of arctic and boveal small
University Press. [JDD] rodents as a problem of production biology. Annales Academiae Scien-
(1971) American social order. New York: Free Press. [JDD] tarum Fennicoe A IV, 66:1-38. [WJH]
Dublin, L. (1963) Suicide. New York: Ronald. [DdC] Kallman, F. J. (1949) Suicide in twins and only children. American Journal of
Dublin, U and Bunzel. B. (1933) To be or not to be. New York: Harrison Human Genetics 1:113-26. [DdC]
Smith and Robert Haas. [DdC] Kaplan, H. B., and Pokorny, A. D. (1976) Self-attitudes and suicidal behavior.
Durham, W. H. (1979) Toward a coevolutionary theory of human biology and Suicide and Life-Threatening Behavior 6:23-25. [DdC, GF]
culture. In Evolutionary biology and human social behavior: an anthro- Kiev, A. (1977) The suicidal patient: recognition and management. Chicago:
pological perspective, ed. N. A. Chagnon and W. Irons. North Scituate, Nelson-Hall. [DdC]
Mass.: Duxbury Press. [DdC] King, M. C , and Wilson, A. C. (1975a) Evolution at two levels in humans and
Durkheim, E. (1951)Sufride. Cleneoe, III.: Free Press. [DdC, LDH] chimpanzees. Science 188:107-16. [DdC]
Farber, M. L. (1968) Theory of suicide. New York: Funk & Wagnalls. [DdC, (1975b) Our close cousin, the chimpanzee. New Scientist 67:16-18. [DdC]
MI.F] Kreitman, N. (1977) Parasuicide. New York: Wiley. [DdC]
(1977) Factors determining the incidence of suicide within families. Suicide Kvetnansky, R.; Mitro, A.; Palkovits, M.; Brownstein, M.; Torda, T.; Vigas, M.;
and Life-Threatening Behavior 7:3-6. [DdC] and Mikulaj, L. (1976) Catecholamines in individual hypothalamic nuclei
Farberow, N. L. (1972) Cultural history of suicide. In Suicide and attempted in stressed rats. In Catecholamines and stress, ed. E. Usdin, R. Kvetnan-
suicide, ed. J. Waldenstrom, T. Larsson, and N. Ljungstedt. Stockholm: sky, and I. J. Kopin. Oxford: Pergamon Press. [HA]
Nordiska Bokhandelns Forlag. [DdC] Laing, R. D. (1967) The politics of experience. New York: Ballantine. [DL]
ed. (1975) Suicide in different cultures. Baltimore: University Park Leese, S. M. (1969) Suicide behavior in twenty adolescents. British Journal of
Press. [DdC] Psychiatry 115:479-80. [DdC]
Ferrence, Ft. C ; Jarvis, G. K.; Johnson, F. C ; and Whitehead, P. C. (1975) The Lesch, M., and Nyhan, W. L. (1964) A familial disorder of uric acid metabo-
self-injury patient: sociological and medical profiles and implications for lism and central nervous system function. American Journal of Medicine
prevention. In Self-destructive behavior, ed. A. R. Roberts. Springfield, 36:561. [DdC]
III.: Charles C.Thomas. [DdC] Lester, D. (1968) Note on the inheritance of suicide. Psychological Reports
Freud, S. (1924) Beyond the pleasure principle. New York: Boni and Live- 22:320. [DdC]
right. [DdC] (1972) Why people kill themselves. Springfield, III.: Charles C. Thom-
Frieden, C. (1978) Masochistic behavior as an outcome of role confusion. Ph.D. as. [DdC, MPD, DL]
dissertation. University of Southern California. [GF] (1977) Bereavement and suicide: a positive perspective. In Suicide and be-
Frieden, G., and Fraser, S. C. (1978) Inducing masochistic behavior: the pain reavement, ed. B. L. Danto and A. H. Kutscher, pp. 194-98. New York:
of feeling low. Paper presented at a meeting of the Western Psychological MSS. [DL]
Association, San Francisco. [CF] Levenson, M. (1974) Cognitive correlates of suicidal risk. In Psychological as-
Gallistel, C. R. (1973) Self-stimulation: the neurophysiology of reward and mo- sessment of suicidal risk, ed. C Neuringer. Springfield, 111.: Charles C.
tivation. In The physiological basis of memory, ed. J. A. Deutsch. New Thomas. [DdC, RDW]
York: Academic Press. [DdC] Levine, S. (1976) Stress and behavior. In Progress in psychobiology, ed. R. F.
Ganzler, S. (1967) Some interpersonal and social dimensions of suicidal behav- Thompson. San Francisco: W. H. Freeman and Company. [DdC]
ior. Dissertation Abstracts 28B: 1192-93. [DdC] Levins, R. (1968) Evolution in changing environments. Princeton, N.J.:
Gershon, E. S. (1978) The search for genetic markers in affective disorders. In Princeton University Press. [DSW]
Psychopharmacology: a generation of progress, ed. M. A. Lipton, A. Di- Linden, L. L., and Breed, W. (1976) The demographic epidemiology of sui-
Mascio, and K. F. Killam. New York: Raven Press. [HA] cide. In Suiddology: contemporary developments, ed. E. S. Shneidman.
Gershon, E. S.; Bunney, W. E., Jr.; Leckman, J. F.; Van Eerdewegh, M.; and New York: Grune & Stratton. [DdC]
DeBauche, B. A. (1976) The inheritance of affective disorders: a review of Lingens, E. (1972) Suicide in extreme situations. In Suicide and attempted sui-
data and of hypotheses. Behavior Cenctics 6:227-61. [DdC] cide, ed. J. Waldenstrom, T. Larsson, and N. Ljungstedt. Stockholm: Nor-
Ghiselin, M. T. (1974) The economy of nature and the evolution of sex. Berke- diska Bokhandelns Forlag. [DdC]
ley: University of California Press. [DSW] McCulloch, J. W., and Philip, A. E. (1972) Suicidal behavior. New York: Per-
Goodall, J. (1979) Life and death at Gombe. National Geographic 155:592- gamon Press. [DdC]
621. [DdC] Maynard Smith, J. (1964) Group selection and kin selection. Nature 20:1145-
Gorzalka, B. B., and Mogenson, G. J. (1977) Sexual behavior. In The neurobi- 47. [DdC]
ology of behavior: an introduction, ed. G. J. Mogenson. Hillsdale, N.J.: (1978) The evolution of behavior. Scientific American 239:176-92. [DdC]
Lawrence Erlbaum Associates. [DdC] Mayr, E. (1963) Populations, species, and evolution. Cambridge, Mass.: Har-
Hamilton, W. D. (1964) The genetical evolution of social behavior, journal of vard University Press. [DSW]
Theoretical Biology 7:1-16. [DdC] Medawar, P. B. (1957) The uniqueness of the individual. London: Me-
(1970) Selfish and spiteful behaviour in an evolutionary model. Nature thuen. [DdC]
228:1218-20. [DdC] Menninger, K. (1938) Man against himself. New York: Harcourt, Brace, and
(1971) Selection of selfish and altruistic behavior in some extreme models. In World. [DdC]
Man and lx*ast: comparative social behavior, ed. J. F. Eisenberg and W. Minkoff, K., Bergman, E., Beck, A. T., and Beck, R. (1973) Hopelessness, de-
S. Dillon. Washington, D C : Smithsonian Institution. [DdC] pression, and attempted suicide. American Journal of Psychiatry
Hankoff, L. D. (1961) An epidemic of attempted suicide. Comprehensive Psy- 130:455-459. [RDW]
chiatry 2:294-98. [DdC] Modigh, K. (1976) Influence of social stress on brain catecholamine mecha-
Headley, L. (1975) Jewish suicide in Israel. In Suicide in different cultures, nisms. In Catecholamines and stress, ed. E. Usdin, R. Kvetnansky and I. J.
ed. N. L. Farberow. Baltimore: University Park Press. [DdC] Kopin. Oxford: Pergamon Press. [HA]
Hellon, C. P., and Solomon, M. I. (in press) Suicide and age in Alberta, 1951- Mogenson, G. J., and Phillips, A. G. (1976) Motivation: a psychological con-
1977: a cohort analysis. Archives of General Psychiatry. [RDW] struct in search of a physiological substrate. In Progress in psychobiology
Hendin, W. (1976). Growing up dead: student suicide. In Suiddology: contem- and physiological psychology, vol. 6, ed. J. M. Sprague and A. N. Epstein.
porary developments, ed. E. S. Shneidman. New York: Grune & Strat- New York: Academic Press. [DdC]
ton. | DdC] Murphy, D. L.; Campbell, I.; and Costa, J. L. (1978) Current status of the in-
Henry, A. F., and Short, J. F. (1954) Suicide and homicide. New York: Free doleamine hypothesis of the affective disorders. In Psychopharmacology:
Press. [DL] a generation of progress, ed. M. A. Lipton, A. DiMascio, and K. F. Kil-
Herskowitz, I. H. (1973) Principles of genetics. New York: Macmil- lam. New York: Raven Press. [HA, DdC]
lan. [DdC] Murphy, G. E., and Wetzel, R. D. (in press) Suicide risk by birth cohort in the
Hinde, R. A., and Stevenson-Hinde, J., eds. (1973) Constraints on learning. United States, 1949-1974. Archives of General Psychiatry. [RDW]
New York: Academic Press. [DdC] Neuringer, C. (1962) Methodological problems in suicide research. Journal of
Irons, W. (1979) Natural selection, adaptation, and human social behavior. In Consulting Psychology 26:273-78. [DdC]
Evolutionary biology and human social behavior: an anthropological per- (1968) Divergencies between attitudes toward life and death among suicidal
spective, ed. N. A. Chagnon and W. Irons. North Scituate, Mass.: Duxbury psychosomatic and normal hospital patients. Journal of Consulting and
Press. [DdC] Clinical Psychology 32:59-63. [RDW]
Jacobs, J. (1971) Adolescent suicide. New York: Wiley. [DdC] (1973) Attitude toward self in suicidal individuals. Life-Threatening Behav-
Jonas, A. D., and Jonas, D. F. (1975) An evolutionary context for schizophre- ior 4:96-106. [RDW]
nia. Schizophrenia Bulletin 2:33-41. [DL] (1974) Self and other appraisals by suicidal psychosomatic and normal hospi-

talized patients. Journal of Consulting Clinical Psychology Books. [DdC, RDW]
42:306. [RDW] Stone, E. A. (1975) Stress and cateeholamines. In Cateeholamines and behav-
(1976) Current developments in the study of suicidal thinking. In Suicidolo- ior, vol. 2, ed. A. J. Friedhoff. New York: Plenum Press. [HA]
gy: contemporary developments, ed. E. S. Shneidman. New York: Grune Stretch, R.; Orloff, E. R.; and Dalrymple, S. D. (1968) Maintenance of re-
& Stratton. [DdC, RDW] sponding byfixed-intervalschedule of electric shock presentation in squir-
Neuringer, C, and Lettieri, D. J. (1971) Cognitions, attitudes and affect in sui- rel monkeys. Science 162:583-86. [GF]
cidal individuals. Life-Threatening Behavior 1:106-24. [RDW] Tabachnick, N. (1970) The crisis treatment of suicide. California Medicine
Nevin, J. A., and Reynolds, G. S., eds. (1973) The study of behavior. Glenview, 112(6): 1-8. [DL]
III.: Scott, Foresman, and Company. [DdCJ Temoche, A.; Pugh, T. F.; and MacMahon, B. (1964) Suicide rates among cur-
Osmond, H., and Hoffer, A. (n.d.) A comprehensive theory of schizophre- rent and former mental institution patients. Journal of Nervous and Men-
nia. [DL] tal Disorders 138:124-30. [DdC]
Otto, U. (1972) Suicidal behaviour in childhood and adolescence. In Suicide Toolman, J. M. (1968) Suicide in childhood and adolescence. In Suiddal be-
and attempted suicide, ed. J. Waldenstrom, T. Larsson, and N. Ljung- haviors, ed. H. L. P. Resnik. Boston: Little Brown and Co. [DdC]
stedt. Stockholm: Nordiska Bokhandelns Forlag. [DdC] Tripodes, P. (1976) Reasoning patterns in suicide notes. In Suiddology: con-
Pare, C. M. B., and Mack, J. W. (1971) Differentiation of two genetically spe- temporary developments, ed. E. S. Shneidman. New York: Grune & Strat-
cific types of depression by the response to anti-depressants. Journal of ton. [DdC)
Medical Cenetics 8:306. [DCB] Trivers, R. L. (1971) The evolution of reciprocal altruism. Quarterly Review of
Parker, C. E. (1978) Selfish genes, evolutionary games, and the adaptiveness of Biology 46:35-57. [DdC]
behaviour. Nature 274:849-55. [DdC] (1972) Parental investment and sexual selection. In Sexual selection and the
Phillips, D. P. (1974) The influence of suggestion on suicide: substantive and descent of man, ed. B. Campbell. Chicago: Aldine. [WJH]
theoretical implications of the Werther effect. American Sociological Re- Venkoba Rao, A. (1975) Suicide in India. In Suidde in different cultures, ed.
view 39:340-54. [DdC] N. L. Farberow. Baltimore: University Park Press. [DdC]
Pokorny, A. D. (1974) A scheme for classifying suicidal behaviors. In The pre- Wade, M. J. (1978) A critical review of the models of group selection. Quar-
diction of suicide, ed. A. T. Beck, H. L. P. Resnik, and D. J. Lettieri. terly Review of Biology 53:101-14. [DdC]
Bowie, Md.: Charles Press. [DdC] Washburn, S. L. (1978) Human behavior and the behavior of other animals.
Renner, K. E., and Tinsley, J. B. (1976) Self-punitive behavior. In The psychol- American Psychologist 33:405-18. [DdC]
ogy of learning and motivation - advances in research and theory, ed. G. Weil-Maherbe, H. (1972) The biochemistry of affective disorders. In Hand-
H. Bower, pp. 155-98. New York: Academic Press. [GF] book of neurochemistry, vol. 7, ed. A. Lathja. New York: Plenum
Richman, J. (1971) Family determinants of suicide potential. In Identifying Press. [DdC]
suicide potential, ed. D. B. Anderson and L. J. McClean. New York: Be- Weiss, J. M. A. (1968) Suicide in the aged. In Suicidal behaviors, ed. H. L. P.
havioral Publications. [DdC] Resnik. Boston: Little Brown and Co. [DdC]
Ridley, M., and Dawkins, R. (in press) The natural selection of altruism. In Al- Weiss, J. M.; Glazer, H. I.; and Pohorecky, L. A. (1976) Coping behavior and
truism and helping behaviour, ed. J. P. Rushton and R. M. Sorrentino. neurochemical changes: an alternative explanation for the original
Hillsdale, N.J.: Lawrence Erlbaum Associates. [RD] "Learned Helplessness" experiments. In Animal models in human psy-
Ritter, S., and Pelzer, N. L. (1978) Magnitude of stress-induced brain norepi- chobiology, ed. G. Serban and A. Kling. New York: Plenum Press. [HA]
nephrine depletion varies with age. Brain Research 152:170. [HA] Welch, B. L., and Welch, A. S. (1970) Control of brain cateeholamines and se-
Roberts, A. R., ed. (1975) Self-destructive behavior. Springfield, 111.: Charles C. rotonin during acute stress and after d-amphetamine by natural inhibition
Thomas. [DdC] of monoamine oxidase: an hypothesis. In Amphetamines and related
Sainsbury, P. (1955) Suicide in London. London: Chapman & Hall. [DdC, compounds, ed. E. Costa and S. Garattini. New York: Raven Press. [HA]
RDW] Wetzel, R. D. (1975a) Ratings of life and death and suicide intent. Psychologi-
Schildkraut, J. J. (1978) Current status of the catecholamine hypothesis of af- cal Reports 27:879-85. [RDW]
fective disorders. In Psychopharmacology: a generation of progress, ed. (1975b) Self-concept and suicide intent. Psychological Reports 36:279-
M. A. Lipton, A. DiMascio and K. F. Killam. New York: Raven 82. [RDW]
Press. [HA] (1976a) Hopelessness, depression, and suicide intent. Archives of General
Schildkraut, J. J., and Kety, S. S. (1967) Biogenic amines and emotion. Science Psychiatry 33:1069-73. [RDW]
156:21-30. [DdC] (1976b) Semantic differential ratings of concepts and suicide intent. Journal
Schneider, E. V. (1953) Sociological concepts and psychiatric research. In In- of Clinical Psychology 32:4-12. [RDW]
terrelations between the social environment and psychiatric disorder. Wetzel, R. D., and Murphy, G. E. (in preparation) A model to predict future
New York: Milbank Memorial Fund. [DdC] suicide rates in the United States. [RDW]
Selye, H. (1973) The evolution of the stress concept. American Scientist Wetzel, R. D., Margulies, T., Davis, R., and Karam, E. (in press) Depression,
61:692-99. [DdC] hopelessness and suicide intent. Journal of Clinical Psychiatry. [RDW]
Shaffer, D. (1974) Suicide in childhood and early adolescence. Journal of Wiens, J. A. (1971) On group selection and Wynne-Edwards' hypothesis. In
Child Psychology and Psychiatry 15:275-91. [DdC] Natural regulation of animal populations, ed. I. A. McClaren. New York:
Shapiro, A. M. (1975) Beau Geste? American Naturalist 109:900-02. [WJH] Atherton Press. [DdC]
Shneidman, E. S. (1957) The logic of suicide. In Clues to suicide, ed. E. S. Williams, G. C. (1957) Pleiotropy, natural selection, and the evolution of sene-
Shneidman and N. L. Farberow. New York: McGraw-Hill. [DdC] scence. Evolution 11:398-411. [DdC, RD]
(1968) Classifications of suicidal phenomena. Bulletin of Suiddology 2:1- Wilson, D. S. (1980) The natural selection of populations and communities.
9. [DdC] Menlo Park, Calif.: Benjamin/Cummings. [DSW]
(1969) Prologue. In On the nature of suicide. San Francisco: Jossey-Bass. Wilson, E. O. (1971) The insect sodeties. Cambridge, Mass.: Belknap. [DdC]
(1976a) Current overview of suicide. In Suiddology: contemporary develop- (1975) Sodobiology. Cambridge, Mass.: Belknap. [DdC, JDD, WJH]
ments, ed. E. S. Shneidman. New York: Grune & Stratton. [DdC] (1978) On Human Nature, Cambridge, Mass.: Harvard University
(1976b) Suicide among the gifted. In Suiddology: contemporary develop- Press. [JDD]
ments, ed. E. S. Shneidman. New York: Grune & Stratton. [DdC] Winokur, G. (1978) Mania and depression: family studies and genetics in rela-
(1976c) Suicide notes reconsidered. In Suiddology: contemporary develop- tion to treatment. In Psychopharmacology: a generation of progress, ed.
ments, ed. E. S. Shneidman. New York: Grune & Stratton. [DdC] M. A. Lipton, A. DiMascio, and K. F. Killam. New York: Raven
Skinner, B. F. (1969) Contingencies of reinforcement: a theoretical analysis. Press. [HA]
New York: Appleton-Century-Crofts. [DdC, WJH] Worden, J. W. (1976) Lethality factors and the suicide attempt. In Suiddolo-
Slater, M. K. (1959) Ecological factors in the origin of incest. American An- gy: contemporary developments, ed. E. S. Shneidman. New York:
thropologist 61:1042-59. [DL] Grune & Stratton. [DdC]
Stanley, S. M. (1975) A theory of evolution above the species level. Proceedings Wright, S. (1922) Coefficients of interbreeding and relationship. American
of the National Academy of Sciences 72:646-50. [DdC] Naturalist 56:330-39. [DdC]
Stein, L. (1978) Reward transmitters: cateeholamines and opioid peptides. In Wynne-Edwards, V. C. (1962) Animal dispersion in relation to social behav-
Psychopharmacology: a generation of progress, ed. M. A. Lipton, A. Di- iour. New York: Hafner. [DdC]
Mascio, and K. F. Killam. New York: Raven Press. [DdC] (1971) Self-regulating systems in populations of animals. In Natural regula-
Stearns, S. C. (1976) Life history tactics, a review of the ideas. Quarterly Re- tion of animal populations, ed. I. A. McClaren. New York: Atherton
view of Biology 51:3-47. [DSW] Press. [DdC]
Stengel, E. (1973) Suicide and attempted suicide. Harmondsworth: Penguin

Baechlers Theory of Suicide

Uploaded by

Copyright:

Available Formats

Baechlers Theory of Suicide

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Baechlers Theory of Suicide

Uploaded by

Copyright:

Available Formats

THE BEHAVIORAL AND BRAIN SCIENCES (1980), 3,265-290

Printed in the United States ol America

Keywords: suicide; self-destruction; sociobiology; evolution; altruism; life-threatening behavior

91980 Cambridge University Press 0140-525X/80/020265-26$04.00/0 265

https://doi.org/10.1017/S0140525X00004799 Published online by Cambridge University Press

266 THE BEHAVIORAL AND BRAIN SCIENCES (1980), 3

https://doi.org/10.1017/S0140525X00004799 Published online by Cambridge University Press

THE BEHAVIORAL AND BRAIN SCIENCES (1980), 3 267

https://doi.org/10.1017/S0140525X00004799 Published online by Cambridge University Press

268 THE BEHAVIORAL AND BRAIN SCIENCES (1980), 3

https://doi.org/10.1017/S0140525X00004799 Published online by Cambridge University Press

THE BEHAVIORAL AND BRAIN SCIENCES (1980), 3 269

https://doi.org/10.1017/S0140525X00004799 Published online by Cambridge University Press

270 THE BEHAVIORAL AND BRAIN SCIENCES (1980), 3

https://doi.org/10.1017/S0140525X00004799 Published online by Cambridge University Press

THE BEHAVIORAL AND BRAIN SCIENCES (1980), 3 271

https://doi.org/10.1017/S0140525X00004799 Published online by Cambridge University Press

272 THE BEHAVIORAL AND BRAIN SCIENCES (1980), 3

https://doi.org/10.1017/S0140525X00004799 Published online by Cambridge University Press

THE BEHAVIORAL AND BRAIN SCIENCES (1980), 3 273

https://doi.org/10.1017/S0140525X00004799 Published online by Cambridge University Press

274 THE BEHAVIORAL AND BRAIN SCIENCES (1980), 3

https://doi.org/10.1017/S0140525X00004799 Published online by Cambridge University Press

THE BEHAVIORAL AND BRAIN SCIENCES (1980), 3 275

https://doi.org/10.1017/S0140525X00004799 Published online by Cambridge University Press

276 THE BEHAVIORAL AND BRAIN SCIENCES (1980), 3

https://doi.org/10.1017/S0140525X00004799 Published online by Cambridge University Press

THE BEHAVIORAL AND BRAIN SCIENCES (1980), 3 277

https://doi.org/10.1017/S0140525X00004799 Published online by Cambridge University Press

278 THE BEHAVIORAL AND BRAIN SCIENCES (1980), 3

https://doi.org/10.1017/S0140525X00004799 Published online by Cambridge University Press

THE BEHAVIORAL AND BRAIN SCIENCES (1980), 3 279

https://doi.org/10.1017/S0140525X00004799 Published online by Cambridge University Press

280 THE BEHAVIORAL AND BRAIN SCIENCES (1980), 3

https://doi.org/10.1017/S0140525X00004799 Published online by Cambridge University Press

and therefore appropriately compared with other aggressive behaviors

THE BEHAVIORAL AND BRAIN SCIENCES (1980), 3 281

https://doi.org/10.1017/S0140525X00004799 Published online by Cambridge University Press

contribute significantly to the advancement of research on suicide. A

282 THE BEHAVIORAL AND BRAIN SCIENCES (1980), 3

https://doi.org/10.1017/S0140525X00004799 Published online by Cambridge University Press

THE BEHAVIORAL AND BRAIN SCIENCES (1980), 3 283

https://doi.org/10.1017/S0140525X00004799 Published online by Cambridge University Press

284 THE BEHAVIORAL AND BRAIN SCIENCES (1980), 3

https://doi.org/10.1017/S0140525X00004799 Published online by Cambridge University Press

THE BEHAVIORAL AND BRAIN SCIENCES (1980), 3 285

https://doi.org/10.1017/S0140525X00004799 Published online by Cambridge University Press

286 THE BEHAVIORAL AND BRAIN SCIENCES (1980), 3

https://doi.org/10.1017/S0140525X00004799 Published online by Cambridge University Press

THE BEHAVIORAL AND BRAIN SCIENCES (1980), 3 287

https://doi.org/10.1017/S0140525X00004799 Published online by Cambridge University Press

288 THE BEHAVIORAL AND BRAIN SCIENCES (1980), 3

https://doi.org/10.1017/S0140525X00004799 Published online by Cambridge University Press

THE BEHAVIORAL AND BRAIN SCIENCES (1980), 3 289

https://doi.org/10.1017/S0140525X00004799 Published online by Cambridge University Press

290 THE BEHAVIORAL AND BRAIN SCIENCES (1980), 3

https://doi.org/10.1017/S0140525X00004799 Published online by Cambridge University Press

You might also like