Suicide in Japan

Decreasing for the living, increasing for the dead

By William Wetherall

A version of this report appeared in three parts as follows:

Part 1: "Suicide in Japan: Figures are misleading"
Decreasing for the living, increasing for the dead
The Japan Times, 19 August 1985, page 6

Part 2: "Suicide in Japan: In the 1980s"
Men who survived the '50s can't cope with the '80s
The Japan Times, 26 August 1985, page 8

Part 3: "Suicide and the individual personality"
Blaming everything but individual personality
The Japan Times, 23 September, page 12

Table (Part 2): Prewar, postwar, and recent rates by sex and age

Suicide in Japan -- Part 1

Suicide in Japan: Figures are misleading
"Decreasing for the living, increasing for the dead"
The Japan Times, 19 August 1985, page 6

Only sociologists are seduced by come-ons like "Care to see my latest statistics?" Anthropologists walk away without even asking what they're about. But even right-brained statisticians don't want to look at suicide rates.

Japan, in any case, is a sociologist's paradise. You name it and someone has counted it. And the data probably go back many years.

Vital (birth, death, marriage, divorce) statistics are available for most of the past century, except for the years 1944-1946. Nationwide divorce rates can be computed annually from 1883. Suicide rates have been reported since at least 1900.

We encounter such numbers everyday in the press. They are often confusing and sometimes suspicious. But we tolerate them. Why?

When we read that X more people killed themselves last year, up Y percent from the year before, we shake our head and wonder what the world is coming to. Surely more death could not be a good thing.

When more means less

According to the Ministry of Health and Welfare statistics reported in June last year [1984], 24,970 (now officially 24,985) Japanese killed themselves in 1983. Because this figure exceeded the "previous record" of 23,641 suicides reported in 1958, the press called the 1983 figures both the "highest" and the "worst" in history.

The logic seems self-evident. We may quibble about the accuracy of the body counts, but most of us will accept as fact that "more people" are committing suicide. And who would consider "more suicide" an improvement?

The problem is that Japan's population in 1958 was only 92.0 million, while in 1983 it was 118.8 million. So there were 25.7 suicides per 100,000 people in 1958 but only 21.0 in 1983.

Hence 1983 was far from being "the worst year for suicide" in Japanese history. Though the body count was higher than in 1958, the increase meant a lower suicide risk for the total population.

In June this year [1985], the Ministry of Health and Welfare similarly reported 24,334 cases of suicide in 1984. This, as one paper put it, "was down 651 from the all-time record in 1983, but still the second highest" (other papers wrote "the second worst"). But the crude suicide rate was only 20.4 per 100,000.

Adjusted rates

Crude rates, though, are heavily biased by population variables like how many people there are in a certain age group. A different picture of suicide emerges when the rates just cited are adjusted for changes in the age (but also sex) structure of the population.

Using the 1935 census population as a standard, the 1958 and 1982 rates become 23.5 and 12.6 per 100,000 people (9 and 28 percent lower than the respective crude rates). The 1873 and 1984 adjusted rates are very roughly 15.0 and 14.5 (29 percent lower than their respective crude rates).

The greater disparity between the crude and adjusted rates for 1982 (and other recent years), compared to 1958, reflects the radical aging in Japan's population structure during the past quarter century. Because suicide rates increase with age, an aging population will produce more cases of suicide, and thus higher crude suicide rates.

The crude suicide rates (but not the adjusted rates) will increase even when the total population and the suicide rates for each age group of the population remain constant. In fact, over most of the years that the population has been steadily increasing, the suicide rates of the elderly have been decreasing.

Adjustments for employment and marital status are not as easily made but are possible. Just as higher ratios of elderly people in the population increases the crude suicide rate, higher ratios of unemployed and divorced people also raise the crude rate -- but not necessarily the adjusted rate, which might even fall.

The selection of the "reference population" that is to serve as a standard for adjusting (and hence comparing) other populations is not without problems of its own. But adjusted rates based on almost any related population are more comparable than crude rates.

This is true for international comparisons also. Assume, for example, that Xen's suicide rate is higher than Japan's. Does this mean that Xenians are more suicidal than Japanese? not necessarily.

If Zen's population has a higher percentage of elderly people, while Japan's suicide rates are higher for the elderly, then Japan's suicide rate would increase were it computed on the basis of Xen's population, while Zen's rate would decrease were it computed on the basis of Japan's population.

In any case, the difference in the two crude rates would shrink or vanish. It might even turn out that Japan's suicide rate is higher than Xen's when adjusted for population differences.

Cause-of-death rates

The major drawback of both crude and adjusted rates, however, is that they are merely ratios of death due to suicide each year per 100,000 living people. As annual population-based rates, they tell us only the probability that a living person will kill themselves during the year, not the probability that the person will eventually kill themselves.

Moreover, such probabilities apply to any person, not every person -- i.e., to a person in general, not to a known person. The probability changes as soon as a person's sex, age, and other traits are known and the person can be associated with a certain group (e.g., unemployed never-married females 25-29 years of age) for which a more specific suicide rate can be estimated.

But far more important in understanding the ultimate impact of suicide on our lives is the percentage of dead who have died because of suicide. That is, what are the chances that a person will die of their own hand?

The 1970 suicide rate in Japan adjusted to the 1935 census was nearly fifty percent lower than the 1920 rate (11.9 vs. 18.8). But suicides in 1970 accounted for nearly three times the percentage of deaths than they did in 1920 (2.21 vs. 0.75).

There were over a thousand more suicides in 1983 than in 1958. But the fewer suicides in 1958 accounted for a slightly greater percentage of all deaths (3.46 percent) than did the greater number of suicides in 1983 (3.37 percent). However, 1983's percentage of suicides as a cause-of-death approached 1958's percentage with only two-thirds as many suicides based on adjusted rates (respectively 15.0 and 23.5).

While suicide rates in Japan during the postwar years have averaged the same or less than those before World War II, the population has doubled while the annual number of deaths from all causes has fallen by half. Against this four-fold reduction in general mortality, suicide and other non-medical causes of death have become more conspicuous.

Suicide and longevity

So suicide as a behavior, while decreasing in proportion to the living population, is increasing in proportion to the dead population. Suicide is more common as a cause of death, not because people are more apt to commit suicide while living, but because they are less likely to die from other causes of death before reaching old age.

Because more people are living longer, the older age groups are swelling in proportion to the younger age groups. And so more people are exposed to the higher suicide risks that attend old age. Hence there are more suicides in both the older age groups and in the total population, but not necessarily higher suicide rates.

This is not intended to be a verbal maze. It is simply a dynamic description of suicide that takes into account the impact of population changes over time. Conventional crude rates are static because they reflect population conditions at only one point in time.

Yet reports like the following composite often appear in the press, and even in academic journals.

4,088 men in their 40s suicided in 1983 (up 35 percent from 3,028 in 1982), while 3,558 men in their 50s killed themselves (up 47 percent from 2,421). This means that 7,646 middle-aged men committed suicide in 1983, or 2,197 (40 percent) more than the previous year, and about three times the 1973 figure.

The picture is greatly exaggerated because population changes were ignored. The number of suicides in the 40-59 age group tripled since 1973. But the 1983 population for the age group grew by nearly 50 percent. So the 1983 rate for the 40-59 age group was only about double (not "three times") the 1973 rate.

False impressions

Journalists are known to misuse numbers when editorializing about alleged social problems. Their impressions of how society is changing are based on official statistics, which are taken at face value. Even experts tend to take the meaning of official statistics for granted. Very few researchers bother to look beneath the surface. many are complacent simply to press the raw published numbers into the service of whatever theories happen to be popular in their academic circles.

Media examples about. A front-page column of the morning edition of one of Japan's national dailies observed, based on National Police Agency statistics, that "4,891 people 65 or older killed themselves in 1978, while in 1983 the number of such suicides was 5,572; as ever, there are many suicides among the elderly."

The editorial went on to discuss the "conspicuous trend toward rushing to death" and attributed "the worst figures in the postwar period" to society's failure to save distressed people who really want to live. And what true patriot wouldn't agree with the implication that postwar Japan is going to the dogs?

The "cry for help" would have been less credible had the article pointed out that the 65-and-over population had grown from 9.9 to 11.7 million during the six-year period mentioned, and that the 1983 suicide rate for the elderly (47.7) was therefore lower than the 1978 rate (49.5).

To be sure, the 1983 rate was higher than the 1982 rate (42.4) based on Ministry of Health and Welfare figures. But the increase still falls far short of the elderly rates in the quarter century following the war, during which the rates were much lower than those of the quarter century before the war. So why all the fuss?

Suicide in Japan -- Part 2

Suicide in Japan: In the 1980s
"Men who survived the '50s can't cope with the '80s"
The Japan Times, 26 August 1985, page 8

The age group trends in Japanese suicide rates over the past three-fourths of a century are clear (see Table). Suicide rates for most age groups during the postwar period 1947-1972 averaged much less than those for the prewar period 1918-1943.

The averages over the recent decade 1973-1982 tend to be even lower -- except for middle-aged men, and young adolescents of both sexes.

Since 1945, suicide has decreased among the younger and older age groups of both sexes, but particularly among younger women. The rates for the younger groups soared to world highs in the 1950s, but just as quickly subsided to more moderate levels in the 1960s.

Gone are the precipitous peaks in the suicide rate curves for young adults in their 20s, and the steep slopes for adolescents in their late teens.

Early-adolescent (10-14) rates increased in the 1970s, but then fell without nearing prewar records. Late-adolescent (15-19) and early-young-adult (20-24) rates sharply decreased over the years of increased competition on exams for entering colleges and companies. But the postwar rates for both adolescents and young adults (20-29) have dropped much more for females than for males.

While the general impression is that Japan's "examination hell" is a seedbed of self-destruction, neither statistics nor case studies support a conclusion that entrance exams are major causes of suicide among the young

Even in cases where the ostensible "precipitating cause" is real or anticipated failure in an exam, the "underlying causes" have been personal or interpersonal factors that seem very remote from the education system as such.

For some young people, deemphasizing exams might remove exams from the long list of contributing factors that can trigger suicide. But there is no evidence that exam-aged youth as a whole would commit less suicide simply because they no longer had to face exams in competition for schools and jobs.

Death-cycle changing

Between 1918 and the mid-1970s, the suicide rates for the middle-aged (30-59) of both sexes remained relatively low and fairly constant, while the rates for youth and the elderly -- but especially adolescents and young adults--moved up and down, sometimes wildly.

So Japan's suicide curves, in which suicide rates are plotted against age, have traditionally shown a deep middle-aged valley between the younger and older mountains on either side.

But about ten years ago, the rates for middle-aged men began to increase, first for men in their late 30s, then for those in their early and late 40s, and now for those in their early and late 50s.

Women continue to face relatively lower suicide risks in the middle years of their life cycle, although recent rates have been slightly higher than the average for the postwar years.

In Japan as in many other countries, suicide rates increase with age. But the postwar rates for the elderly (60 and over) are not nearly as high as the prewar rates (see Part 1). Despite the rumored ill effects of family nuclearization on the mental health of the aged, the incidence of suicide among the elderly of both sexes -- but especially men -- has tended to decline.

The only notable exception to the long-term trend towards lower suicide rates in Japan are men in their 40s and 50s, who indeed are taking their own lives at unprecedented rates.

Among men in the middle decades of their life, the 1983 rate was highest for the 50-54 age group -- 55.6 per 100,000 -- which broke the 54.3 record set at the height of the depression in 1932.

The 50-56 age-group averaged 29.9 for the decade 1973-1982, during which it steadily increased from 22.7 to 38.6. This was only slightly lower than the postwar 1947-1972 average (29.9), though much lower than the prewar 1918-1943 average (42.1).

The currently elevated rates for middle-aged men will have to remain high for many years before the overall postwar average exceeds the prewar average.

Middle-aged men

The trend toward higher rates of suicide among middle-aged men is not new. The past two years have marked only the most violent eruption of the volcano that began to emerge about a decade ago in what was once a relatively peaceful valley in the middle of the male life cycle -- a haven for men who were then more certain of their raison d'etre in Japanese society, and more secure in their jobs and marriages.

This "peaceful valley" understanding of Japan's U-shaped suicide curve is, however, an optical illusion. When the curve is factored by marital status (i.e., when the total curve is broken down into four sub-curves representing never-married, married, widowed, and divorced), the real meaning of the peak and valley become apparent, and the U-ness of the general curve loses its meaning.

Many (but hardly all) of the men who are facing the greater suicide risks in their middle years are members of the "one-digit Showa" generation. Such men were born between 1926 and 1935, the first decade of the Showa era [the era of the present emperor, Hirohito, whose reign name is Showa] (1984 is Showa 59).

Some critics are blaming the increase in the suicide rate among middle-aged men on the social environment which is though to have shaped their characters. Middle managers in their 40s and 50s are said to be stressfully sandwiched between the world views of the authoritarian older managers who have dominated their careers, and the carefree younger workers who they in turn must supervise.

Burdens heavier

A government white paper on livelihood, published in November 1984, reported results of an attempt to analyze the economic life cycle of the average bread winner -- a married man with two children. The model-household male between 48 and 55 has to spend far more than he makes in order to manage home mortgages and the costs of his children's educations.

He must bear this burden at the height of his career, when his company responsibilities are heaviest, and his leisure time is most limited. He is also at a stage in his life when he becomes increasingly aware that his physical strength is beginning to ebb, and that mentally he is aging faster than he wants to believe.

But the middle-aged men who are rushing to death have a multitude of problems, not all of which are attributable to the demands of corporate and family life, or to psychological reactions to biological changes. The great majority of their contemporaries have managed to endure the same hardships -- long and arduous commuting journeys overtime and drinking obligations after regular hours, short vacations, assignments and transfers which separate them from their families, and intense rivalry for promotions.

The fact that the victims may be divorced or have domestic problems, or be in debt over their heads, or feel that they cannot keep up with new technology ("technostress"), suggests that their grief is also caused by personal factors that cannot be easily blamed on society or culture.

Early-Showa men

The generation of the early-Showa is paradoxically praised for its unselfish diligence in sustaining Japan's high growth rate through the critical postwar years. A social Darwinist would argue that this generation's higher suicide rates reflect its collective ruling against itself, that it is no longer useful to the country it helped rebuild.

The life-employment and seniority systems, at least in the larger companies and in the national and local civil services, promised to compensate him for the sacrifices he made on the altar of national achievement with rewards of higher status and pay when older. But the job security carrot began to go limp about the time of the oil crises in the 1970s when employers discovered that they could no longer afford to pay more and more for the services of workers who become less and less productive with age.

Longer life expectancies and later retirement ages, together with lower birth rates (hence shortages of younger workers), have increased both the number and percentage of older men in the work force, thus further undermining the seniority system.

"The early-Showa generation lost its life goals when Japan was defeated," said Hamamatsu University psychiatrist Kenshiro Ohara in a magazine interview last year. "Everything they were told by their parents and older siblings took a 180 degree turn. When humans lose what they consider most important, they commit suicide."

Ohara predicted that this generation may face even higher suicide rates as it moves into retirement, because many of its members have never learned to relax and enjoy life.

Suicide Generation

But other common experiences in childhood and adolescence may have exacerbated the early-Showa generation's propensity for suicide. The very same generation was decimated by suicide in the 1950s when its members were in their late teens and 20s.

So when they were in the prime of their youth, those who are committing suicide today lost many of their age peers to "the sickness unto death" caused mainly by depression, chronic and acute. Some of them also lost their parents, siblings, and friends during the war.

Such early death experiences may even go back to the late 1920s and the 1930s, when their parents' generation was in its late teens and twenties, and were setting prewar suicide records of its own in the midst of economic calamity and rising militarism.

"[Early-Showa men] have worked hard and become affluent," said psychiatrist Ohara, well-known for his studies of suicide. "But they lose their object in living when, as middle managers, their diligence goes unappreciated by those above and below them."

While this suggests a process of demographic thinning within a generation that finds itself "used" and discarded, suicide remains a very complex and enigmatic act.

The social and cultural "causes" and "reasons" reported in journalistic and even academic accounts of an individual act of self-destruction may well have been contributing or even precipitating factors. Case histories, though, tend to reflect psychological variations that belie attempts to reduce suicide in to a purely social or cultural behavior.

Suicide in Japan -- Part 2

Prewar, postwar, and recent rates by sex and age

(averages for prewar, postwar, and recent periods)

Period  /  Age Group   5-9  10-14  15-19  20-24  25-39  30-34  35-39  40-44  45-49  50-54  55-59  60-64  65-69  70-74  75-79  80-84  85 +


A   Prewar   1918-1943  .00  1.46  20.1   41.6   32.6   24.3   25.3   27.2   34.2   42.1   51.1   68.4 ------> 103.6 ------> 134.3 ------>

B   Postwar  1947-1972  .03   .93  17.8   42.0   33.9   21.9   19.3   19.2   24.0   29.9   39.8   48.2   61.5   75.6   93.9  113 *  125 *

C   Recent   1973-1982  .02  1.35  11.3   25.0   25.6   23.7   25.4   28.9   30.5   29.3   30.3   34.0   39.2   50.9   68.1   85.6  100.2

% change from A to B      +   -36   -11     +1     +4    -10    -24    -29    -30    -29    -22    -30      -    -27      -    -16      -

% change from B to C    -33   +45   -37    -40    -24     +8    +32    +51    +27     -2    -24    -29    -36    -33    -27    -24    -20


A   Prewar   1918-1943  .00  1.02  19.7   27.4   19.4   16.8   15.4   15.7   17.7   19.7   21.5   31.8 ------>  57.3 ------>  80.9 ------>

B   Postwar  1947-1972  .01   .42  13.9   29.3   20.3   14.7   12.6   12.4   14.7   17.4   20.8   28.0   40.0   53.8   67.5   80 *   86 *

C   Recent   1973-1982  .01   .53   6.1   13.8   13.4   12.3   12.7   12.3   13.7   16.6   18.1   22.2   31.7   44.7   60.6   70.6   71.4

% change from A to B      +   -59   -29     +7     +5    -13    -18    -21    -17    -12     -3    -12      -     -6      -     -1      -

% change from B to C      0   +26   -56    -53    -34    -16     +1     -1     -7     -5    -13    -21    -21    -17    -10    -12    -17
 *  Rates estimated from data for all years except 1951-1954

Suicide in Japan -- Part 3

Suicide and the individual personality
"Blaming everything but individual personality"
The Japan Times, 23 September, page 12

"Suicide in Japan is widely misunderstood," warned Yasuhiko Yuzawa, a professor at Ochanomizu University in Tokyo [who specializes in family problems], in a magazine interview last year.

"Many people are under the impression that suicide among the elderly has been increasing during the postwar period because the traditional family system and respect for the aged have been lost. But the prewar rates are much higher than the postwar rates."

Such views fly in the face of a politically influential 1981 study which criticized junior high school textbooks that positively describe the postwar trend toward the nuclear (two-generation) family, away from the traditional extended family. The study concluded that failure to mention the increasing number of elderly who are forced to live alone is reflected in the fact that the suicide rates for Japan's elderly are among the highest in the world.

Editorials on the study failed to note that suicide rates among the aged in Japan, though still high in comparison with many other countries, have declined during the period that nuclear families have increased (See Parts 1 and 2).

The long-term trend toward lower suicide rates among the elderly, despite occasional flares, does not permit the conclusion that older people who live alone are less suicide prone than age peers who live with their children. It does mean, though, that critics of nuclearization have to come up with other reasons for blaming elderly suicides on the "creeping westernization" of Japan.

Blaming woman

Ideology also colors the reasons commonly given by experts for the increasing suicide rates among middle-aged men. Prominent Japanese mental health specialists generally agree that changes in the home are important contributing factors.

A recent Asahi jaanaru (Asahi Journal) article expressed the prevailing sentiments as follows (9 August 1985, page 11):

For middle-aged men, the home was once a port of calmness. It was a place where, returning from the rough seas of work, a man could say what he wanted and have his own way. Now he goes home to a wife who wants a divorce and to children who rebel. To schematicize this in the extreme, wives and children are killing their husbands and fathers.

Worth nothing here is how the wife is subtly being blamed for her husband's failure to accommodate his changing family. Most Japanese mental health specialists are men, and when talking about suicides among their middle-aged peers they like to joke about women's liberation, and they nervously laugh at how strong and independent women are becoming.

But here, too, blame readily falls on family nuclearization, the idea being that a wife would learn how to titillate her husband's ego if her mother-in-law, who indulged his ego-centrism until his marriage, were there to show her how to maintain his suspended state of prepubescent dependence, and help keep the kids in line too.

The circularity of this argument is compounded by particularistic preconceptions that only Japanese men face such problems. One wonders where the sexual paranoia ends and the nationalistic self-pity begins.

Sandwich syndrome

Suicide preventionist Sohei Akiyama has studied seven cases of suicide that occurred between 1981 and 1983 among Japanese National Railway supervisors in Kyushu. He has dubbed them "sandwich syndrome suicides" because all of the men, as newly-arrived transferees ranging in age from 36 to 53, were caught between pressure and hostility from subordinates and lack of support from supervisors.

Akiyama apparently had trouble getting his point across at the 13th International Congress for Suicide Prevention and Crisis Intervention held in Vienna in early July this year. After Akiyama presented his paper, American suicidologist Edwin S. Shneidman reportedly commented that the expression "sandwich syndrome suicide" would be "difficult for foreigners to understand" because "sandwiches are soft: they are not things that crush people" (Asahi jaanaru, 9 August 1985, page 9).

If this was an effort at humor, it was lost on the Japanese delegates, who returned to Japan convinced (if not before they left) that the "syndrome" is peculiar to middle-aged Japanese men. This may be another case of confusing the culture of suicide (how a society needs to explain suicide) for the behavior of suicide (the personal, universal, human reasons people kill themselves).

So long as behavior is classified in terms of socio-cultural particulars, it will always be possible for Japanese researchers to wallow in the comforts of Nihonjinron -- those theories that hold Japanese to be unique.

But behind the labels that are coined as though to maintain the myth that Japanese kill themselves for different reasons than other people, the universalist will recognize the reactive depression that commonly attends maladjustment in cases of both voluntary and involuntary relocation.

Stress universals

All people who change jobs or are transferred, like all people who move, probably experience anxiety and other psychological reactions. But just as most Japanese sojourners who must live and work in New York do not kill themselves, most job transferees survive. No one is ever "driven" to suicide by "culture shock" or by "sandwiching" between anything.

Some people are more susceptible of suicide because they lack the personality traits and other emotional resources that would enable them to cope with their environments and situations, however different or alienating they may suddenly become.

It is, of course, possible to "prevent" suicide by improving or avoiding the problematical environments and situations. But this is like stopping the growth of a bacterial culture by refrigeration: as soon as the electricity goes off and the temperature rises, the culture will resume growing.

No matter how much interpersonal relations may trigger the "social problem" called suicide, it is the individual -- not the group -- that ultimately commits the act. Dealing with precipitating factors is at best a stopgap measure, because life simply does not guarantee that susceptible individuals will always be protected from unfriendly environments and alienating situations.

Suicide prevention should stress reducing susceptibility, which means that efforts should focus on the psycho-dynamics of the suicidal individual. In this sense, comparative suicide research can best aid prevention by trying to identify and understand the common denominators of human frailty, as well as the socio-cultural numerators that, for a given denominator, determine an individual's suicidal risk.

Nature or nurture?

Suicide rates are governed mainly by life-cycle (socio-cultural) factors, while suicide acts are controlled primarily by psychological (biological) factors. As an epidemiological phenomenon, the shift in the male peak from the younger to the middle years of life can be explained in terms of factors like aging and marital status. But a single case of suicide must be dealt with in terms of the individual's psycho-dynamics.

Ultimately very personal mechanisms determine how an individual copes with unfriendly and alienating signals from society, culture, and self. Acquired and learned abilities to handle disappointment, loss of confidence, and self-doubt are also highly personal in nature.

As a psychological act, then, suicide is a resolution of complex biological (neural and hormonal) factors. Yet collective patterns do emerge that admit some socio-cultural interpretations.

Suicide curves, for example, readily lend themselves to life-cycle explanations. A "suicide curve" is a plot of crude suicide rates by age group. Until recently, Japan's curve has been U-shaped. The same kind of curve is found in many other Asian countries but also in some Euro-American nations.

The U-shape (really J-shape) of Japan's traditional curve seems to reflect life-cycle patterns like stress and anxiety in youth (high rates), security in middle age (low rates), and loss of meaning in old age (higher rates). Recently, though, the youthful female peak has virtually disappeared, while the youthful male peak has turned into a middle-aged hump. But the peaks, if not the hump, are better understood in terms of marital status.

Much has been said about the generational factors behind the trend toward higher suicide rates among middle-aged males. But the numbers suggest that the trend is not limiting itself to people born in the early-Showa period. Moreover, there are signs that middle-aged women may be developing a hump of their own.

Quality control

The premature deaths of the early-Showa generation have already made it the most suicidal in Japan's modern century. The middle-aged bulge is also the most dramatic change in Japanese suicide patterns since nationwide statistics have been compiled.

What lessons can be learned from this development? Can we write it off as an anomaly created by schizophrenic social forces, as a generation sheered by prewar and postwar values?

Perhaps. But the middle-aged suicide crisis might also mean that Japan's commitment to nation-building, at the expense of individual mobility and welfare, is not working well.

As a national phenomenon, Japan's middle-aged suicides can be seen as a warning that devotion and diligence are life-threatening values when held obsessively and compulsively in the name of excessive self-denial and dependence on others.

Japan may have over-produced dedicated but socially dull and culturally rigid businessmen and bureaucrats who lack endo-skeletons, and who fall apart when the exo-skeleton of the family or company fails to support their soft mental tissues.

Maybe it is time to give more attention to the quality of the personalities of those who endeavor to make and sell some of the world's finest products, and build and run some of the most reliable trains. Suicide, though, is not the only index of unhappiness. Nor are lower suicide rates guarantees of a better life.