CHAPTER 3 EGOISTIC SUICIDE (continued)

BUT if religion preserves men from suicide only because and in so far as it is a society, other societies probably have the same effect. From this point of view let us consider the family and political society.

I

If one consults only the absolute figures, unmarried persons seem to commit suicide less than married ones. Thus in France, during the period 1873-78, there were 16,264 suicides of married persons while unmarried persons had only 11,709. The former number is to the second as 132 to 100. As the same proportion appears at other periods and in other countries, certain authors had once taught that marriage and family life multiply the chances of suicide. Certainly, if in accordance with current opinion one regards suicide primarily as an act of despair caused by the difficulties of existence, this opinion has all the appearance of probability. An unmarried person has in fact an easier life than a married one. Does not marriage entail all sorts of burdens and responsibilities? To assure the present and future of a family, are not more privations and sufferings required than to meet the needs of a single person?2 Nevertheless, clear as it seems, this a priori reasoning is quite false and the facts only seem to support it because of being poorly analyzed. The elder Bertillon first established this by an ingenious calculation which we shall reproduce.

Really to appreciate the figures given above, we must remember that a very large number of unmarried persons are less than 16 years old, while all married persons are older. Up to 16 years the tendency to suicide is very slight, due to age, without considering other factors. In France only one or two suicides per million inhabitants are found at this time of life; at the following period there are twenty times as many. The inclusion of many children below 16 among unmarried persons thus unduly reduces the average aptitude of the latter, since the reduction is due to age, not celibacy. If they seem to contribute fewer suicides, it is not because they are unmarried but because many of them are yet immature. So, if one tries to compare the two populations to determine the influence of marital status and that alone, one must rid oneself of this disturbing element and compare with married persons only the unmarried above 16. When this subtraction is made, it appears that between 1863-68 there were on the average 173 suicides in a million unmarried persons above 16 years and 154.5 for a million married persons. The ratio of the first to the second number is that of 112 to 100.

There is thus a certain accretion due to celibacy. But it is much greater than the preceding figures show. Actually, we have assumed that all unmarried persons above 16 years and all married persons were of the same average age. This is not true. The majority of unmarried men in France, exactly 58 per cent, are between 15 and 20 years; the majority of unmarried women, exactly 57 per cent are less than 25 years. The average age of all unmarried men is 26.8, of all unmarried women 28.4. The average age of married persons, on the contrary, is between 40 and 45 years. For both sexes combined, suicide develops according to age as follows:

From 16 to 21 years   45.9 suicides per million inhabitants

From 21 to 30 years   97.9 suicides per million inhabitants

From 31 to 40 years   114.5 suicides per million inhabitants

From 41 to 50 years   164.4 suicides per million inhabitants

These figures refer to the years 1848-57. If age were the only influence, the aptitude of unmarried persons for suicide could not be above 97.9 and that of married persons would be between 114.5 and 164.4, or about 140 suicides per million inhabitants.Suicides of married persons would be to those of unmarried as 100 to 69. The latter would be only two-thirds of the former whereas we know that they are actually more numerous. The effect of family life is thus to reverse the relation. Whereas without the effect of family life married persons should kill themselves half again as often as unmarried by virtue of their age, they do so perceptibly less. Thus marriage may be said to reduce the danger of suicide by about half or, more exactly, non-marriage produces an increase expressed by the proportion 112/69, or 1.6. Thus, if we represent the suicidal tendency of married persons by unity, that of unmarried persons of the same average age must be estimated as 1.6.

The relationships are practically the same in Italy. Due to their age, married persons (years 1873-77) should show 102 suicides per million and the unmarried above 16 years only 77; the first number is to the second as 100 to 75. Actually, married persons commit fewer suicides; they show only 71 cases to 86 of unmarried persons or 100 to 121. The aptitude of the unmarried is thus in the proportion of 121 to 75 for that of married persons, or 1.6, as in France. Similar figures might be obtained in other countries. The rate of married persons is everywhere to some degree below that of unmarried persons, whereas it ought, by virtue of age, to be higher. In Wurttemberg, from 1846 to 1860, these two figures were to one another as 100 to 143, in Prussia from 1873 to 1875 as 100 to 111.

But if, with the data available, this method of calculation is the only one applicable in almost all cases, and if consequently it must be used to establish the general situation, its results can be only roughly approximate. Of course, it suffices to show that non-marriage increases the tendency to suicide; but it gives only a very inexact idea of the extent of this increase. Indeed, to distinguish the influence of age and that of marital status, we have taken as our starting point the relation between the suicide-rate at 30 years and that at 45. Unfortunately, the influence of marital status has already left its own mark on this relation; for the contingent of each of the two ages was calculated for unmarried and married persons taken together. Of course, if the proportion of married and unmarried men were the same at the two periods, as well as that of unmarried and married women, they would compensate each other and the effect of age alone would be apparent. But this is not so. While at 30 unmarried men are slightly more numerous than married men (746,111 for the former, 714,278 for the latter according to the census for 1891), at 45 years, on the contrary, the former are only a slight minority (333,-033 to 1,864,401 married men); it is the same with the other sex. Because of this unequal distribution, their great aptitude for suicide does not produce the same effects in both cases. It increases the former rate much more than the latter. The latter is consequently relatively too slight and the numerical superiority which it would show over the former if age alone were involved is artificially reduced. In other words, the difference as regards suicide, due merely to the fact of age, between the population of from 25 to 30 years and that of from 40 to 45 is certainly greater than appears from this way of figuring. Now the extent of this difference forms almost all the relative immunity of married people. This immunity thus appears less than it is in reality.

This method has caused even greater errors. Thus, to determine the influence of widowhood on suicide, the rate of widowed persons has sometimes merely been compared with that of persons of every marital status of the same average age, or about 65 years. Now a million widowers in 1863-68 showed 628 suicides; a million men aged 65 (of every marital status combined) about 461. From these figures one might judge that at the same age widowed persons kill themselves considerably more often than any other class of the population. In this way the assumption has arisen that widowhood is the most unlucky of all states from the point of view of suicide. Actually, if the population of 65 years does not show more suicides, it is because it is almost entirely composed of married persons (997,198 to 134,238 unmarried). So if this comparison suffices to prove that widowed persons kill themselves more than married persons of the same age, it shows nothing as to their tendency to suicide compared with that of unmarried persons.

In short, when only averages are compared, the facts and their relations to one another appear only approximately. Thus it may very well be true that married persons kill themselves in general less often than unmarried persons, and that nevertheless this proportion may be exceptionally reversed at certain ages; in fact we shall see that this is so. Now these exceptions, possibly instructive for the explanation of the phenomenon, could not be shown by the preceding method. There may also be changes from age to age, which without achieving complete inversion, have nevertheless an importance of their own and which should therefore be shown.

The only way to avoid these difficulties is to determine the rate of each group separately, at each age. Under such conditions one may, for example, compare unmarried persons of from 25 to 30 years with married and widowed persons of the same age and similarly for other periods; the influence of marital status will thus be isolated from all the other influences and all its possible variations will appear. Besides, this is the method which Bertillon first applied to mortality and the marriage rate. Unfortunately, official publications do not contain the necessary data for this comparison. Actually, they show the age of suicides independently of their marital status. The only publication which to our knowledge has followed a different practice is that of the grand-duchy of Oldenburg (including the principalities of Lübeck and Birkenfeld). For the years 1871-85 this publication gives us the distribution of suicides by age for each category of marital status considered separately. But this little State had only 1,369 suicides during these fifteen years. As nothing certain can be concluded from so few cases, we undertook to do the work ourselves for France with the aid of unpublished documents in the possession of the Ministry of Justice. We studied the years 1889, 1890 and 1891. We classified about 25,000 suicides in this way. Not only is such a figure sufficiently important in itself to serve as a basis for induction, but we assured ourselves that there was no need to extend our observations over a longer period. From one year to another the contingent of each age remains approximately the same in each group. There is therefore no need to fix the averages for a greater number of years.

Tables XX and XXI contain these different figures. To make their meaning clearer we have placed for each age, beside the figure expressing the rate for widowed persons and that for married persons, what we call the coefficient of preservation, either of the latter by comparison with the former or of both by comparison with unmarried persons. By this phrase we mean the number showing how many times less frequent suicide is in one group than in another at the same age. Thus, when we say that the coefficient of preservation of husbands of the age of 25 in relation to unmarried men is 3, we mean that if the tendency to suicide of married persons at this time of life is represented by 1, that of unmarried persons the same period must be represented by 3. Of course, when the coefficient of preservation sinks below unity, it really becomes a coefficient of aggravation.

TABLE XX—Grand-duchy of Oldenburg: Suicides Committed, by Each Sex, per 10,000 Inhabitants of Each Age and Marital StatusGroup Throughout the Period 1871-85 *

The laws derived from these tables may be formulated thus:

1. Too early marriages have an aggravating influence on suicide, especially as regards men. This result, to be sure, being calculated from a very small number of cases, should be confirmed; in France, from 15 to 20 years, in the average year barely one suicide is committed among married persons, exactly 1.33. However, as the fact is likewise observed in the grand-duchy of Oldenburg, and even for women, it is probably not accidental. Even the Swedish statistics quoted above, show the same aggravation, at least for the male sex. If, now, for the reasons mentioned, we believe these statistics inexact for the advanced ages, we have no reason to doubt them for the first periods of life, when there are as yet no widowed persons. Besides, the mortality of very young husbands and wives is known to considerably exceed that of unmarried men and women of the same age. A thousand unmarried men between 15 and 20 give 8.9 deaths each year, a thousand married men of the same age, 51 deaths or 473 per cent more. The difference is less for the other sex, 9.9 deaths for wives, 8.3 for unmarried women; the former number is to the second only as 119 to 100. This greater mortality of young married persons is evidently due to social reasons, for if its principal cause were the immaturity of the organism this would be more marked in the female sex, due to the dangers involved in parturition. Thus everything tends to prove that premature marriages bring about a harmful moral state, especially to men.

TABLE XXI—France (1889-1891): Suicides Committed per 1,000,000 Inhabitants of Each Age and Marital Status Group, Average Year

2. From 20 years, married persons of both sexes enjoy a coefficient of preservation in comparison with unmarried persons. It is above that calculated by Bertillon. The figure 1.6 indicated by that observer is a minimum rather than an average.

This coefficient changes with age. It soon reaches a maximum between 25 and 30 years in France, between 30 and 40 in Oldenburg; from then on it decreases till the final period of life when a slight rise sometimes occurs.

3. The coefficient of preservation of married persons by comparison with unmarried persons varies with the sexes. In France it is men who are in the favorable position and the difference between the sexes is considerable; for married men the average is 2.73 while for married women it is only 1.56, or 43 per cent less. But in Oldenburg the opposite is true; the average for women is 2.16 and for men only 1.83. It is to be noted that at the same time the disproportion is less; the second number is only 16 per cent lower than the first. We shall say therefore that the sex enjoying the higher coefficient of preservation in the state of marriage varies from society to society and that the extent of the difference between the rates of the sexes itself varies to the extent that the coefficient of preservation favors the favored sex. In the course of our work we shall encounter facts confirming this law.

4. Widowhood diminishes the coefficient of married persons of each sex, but it rarely eliminates it entirely. Widowed persons kill themselves more often than married persons but generally less than unmarried persons. Their coefficient in certain cases even rises to 1.60 and 1.66. Like that of married persons it changes with age, but following an irregular evolution the law of which cannot be determined.

Just as for married persons, the coefficient of preservation of widowed persons compared with unmarried persons varies with the sex. In France men are in the favored position; their average coefficient is 1.32 while for widows it falls below unity, 0.84, or 37 per cent less. But in Oldenburg women are favored, as in marriage; they have an average coefficient of 1.07, while that of widowers is below unity, 0.89, or 17 per cent less. As in the state of marriage, when it is women who are most favored, the difference between the sexes is less than where men have the advantage. So we may say in the same terms that the sex enjoying the higher coefficient of preservation in the state of widowhood varies from society to society, and that the extent of the difference between the rates of the sexes, itself varies to the extent that the coefficient of preservation favors the favored sex.

Facts being thus determined, let us seek explanations.

II

The immunity enjoyed by married persons can be attributed only to one of the two following causes:

It may be due to the influence of the domestic environment. It would then be the influence of the family which neutralized the suicidal tendency or prevented its outburst.

Or this immunity is due to what may be called matrimonial selection. Marriage in fact does make for some sort of selection among the population at large. Not everyone who wants to, gets married; one has little chance of founding a family successfully without certain qualities of health, fortune and morality. People without them, unless through a conjunction of exceptionally favorable circumstances, are thus involuntarily relegated to the unmarried class which consequently includes the human dregs of the country. The sick, the incurable, the people of too little means or known weakness are found here. Hence, if this part of the population is so far inferior to the other, it naturally proves this inferiority by a higher mortality, a greater criminality, and finally by a stronger suicidal tendency. According to this hypothesis, it would not be the family which was a protection against suicide, crime or sickness; the privileged position of married persons would be theirs simply because only those are admitted to family life who already provide considerable guarantees of physical and moral health.

Bertillon seems to have vacillated between the two explanations and to have admitted both at once. Since then, M. Letourneau, in his Evolution du mariage et de la famille, has categorically chosen the second. He refuses to acknowledge that the undeniable superiority of the married population is a result and proof of the superiority of marital life. He would have judged less precipitously had he observed the facts less hastily.

Of course, it is quite probable that married people generally have a physical and moral constitution somewhat better than that of unmarried persons. Matrimonial selection, however, does not bar all but the elite of the population from wedlock. It is especially doubtful that persons without means and position marry much less than others. As has been noted, they usually have more children than the people with assured incomes. If, then, no forethought limits the imprudent increase of their family, why should it prevent their founding one? Besides, repeated proof will be given below that poverty is not one of the factors on which the social suicide-rate depends. As for the infirm, not merely are infirmities overlooked for many reasons, but it is not at all certain that suicides are most numerous among the infirm. The organic-psychic temperament most predisposing man to kill himself is neurasthenia in all its forms. Now today neurasthenia is rather considered a mark of distinction than a weakness. In our refined societies, enamoured of things intellectual, nervous members constitute almost a nobility. Only the clearly insane are apt to be refused admittance to marriage. This limited exclusion is not enough to account for the extensive immunity of married persons.

Besides these somewhat a priori considerations, numerous facts show that the respective immunity of married and unmarried persons is due to quite other causes.

If it were a result of matrimonial selection, it should grow from the start of this selection, or the age when young men and women begin to marry. At this point, a first difference should be noted which should increase with the progress of selection, or as marriageable persons marry and thus lose contact with the rabble naturally destined to be the class of the permanently unmarried. In short, the maximum should be reached when the good grain is completely separated from the tares, when the whole population admissible to marriage has actually been admitted, when only those are unmarried who are hopelessly committed to this condition by physical or moral inferiority. This maximum should occur between 30 and 40 years of age; few marriages are made later.

Now, the coefficient of preservation actually evolves according to quite another law. At first it is often replaced by a coefficient of aggravation. Very young married persons are more inclined to commit suicide than unmarried ones; this would not be so if their immunity were inherent and inherited. Secondly, the maximum is achieved almost at once. At the earliest age when the privileged position of married persons becomes perceptible (between 20 and 25 years), the coefficient reaches a figure which it is unlikely later to surpass. Now at this period there are only 148,000 married to 1,430,000 unmarried men, and 626,000 married to 1,049,000 unmarried women (in round numbers). The ranks of the unmarried, therefore, at this time include the largest part of the elite which has been thought destined by its hereditary qualities to form later the aristocracy of the married; the difference from the point of view of suicide between the two classes should then be slight, whereas it is already considerable. Likewise, at the next age (between 25 and 30), more than a million of the two million married persons to appear between the ages of 30 and 40 are still unmarried; and yet far from the immunity of the unmarried profiting by this fact, this group cuts the poorest figure then. Never are the two parts of the population so far from one another as regards suicide. On the contrary, between the ages of 30 and 40, when the separation is complete and the married class has about reached its full complement, instead of reaching its height and thus showing that conjugal selection itself has come to a stop, the coefficient of preservation undergoes an abrupt and considerable decline. For men it falls from 3.20 to 2.77; for women the regression is still more pronounced, 1.53 instead of 2.22 or a reduction of 32 per cent.

On the other hand, however this selection is effected, it must occur equally for unmarried women as for unmarried men; for wives are recruited in the same manner as husbands. Thus, if the moral superiority of married persons is merely a result of selection, it should be the same for both sexes, and consequently the immunity from suicide should be the same. Actually, husbands are definitely more protected in France than wives. For the former, the coefficient of preservation rises as high as 3.20, falls only once below 2.04, and usually oscillates about 2.80, while for the latter the maximum does not exceed 2.22 (or at most 2.39) and the minimum is below unity (0.98). Moreover, women in France are closest to men with respect to suicide, in the married state. The share of each sex in suicides, for each category of marital status, for the years 1887-91, follows below:

Thus at each age the share of wives in the suicides of married persons is far higher than that of unmarried women in suicides of unmarried persons. Certainly this is not because a wife is less protected than an unmarried woman; Tables XX and XXI show the contrary. But if women do not lose by marriage, they gain less than men. But if immunity is here so unequal, family life must affect the moral constitution of the two sexes differently. What proves with real finality that the inequality has this origin is that its birth and growth may be observed under the influence of the domestic environment. Indeed, Table XXI shows that in the beginning the coefficient of preservation for the two sexes is hardly different (for women, 2.39 in the 15-20 age-group or 2.00 in the 20-25 age-group; for men, 2.40 in the 20-25 age-group. Then, gradually, the difference increases, at first because the coefficient of married women grows less than that of married men up to the age of the maximum, and then because its decrease is swifter and greater., Thus, the coefficient of preservation evolves in accordance with the prolongation of family-life because it depends on this prolongation.

Still better proof is that the relative situation of the sexes as to the degree of preservation enjoyed by married persons is not the same in all countries. In the grand-duchy of Oldenburg women are the favored sex and we shall find later another case of the same inversion. But on the whole, conjugal selection occurs everywhere in the same way. So it cannot be the essential factor in matrimonial immunity; for how then would opposite results occur in different countries? On the contrary, the family may very well be constituted in two different societies so as to affect the sexes differently. In the constitution of the family group, accordingly, we must find the principal cause of the phenomenon of our study.

But interesting as this result is, it must be further defined; for the family environment consists of different elements. For husband and wife alike the family includes: 1. the wife or husband; 2. the children. Is the salutary effect of the family on the suicidal tendency due to the former or the latter? In other words, the family consists of two different associations: the conjugal group and the family group proper. These two societies have not the same origin, nor the same nature, nor consequently, in all probability, the same effects. One springs from a contract and elective affinity, the other from a natural phenomenon, consanguinity; the former unites two members of the same generation, the latter unites one generation to the next; the latter is as old as humanity, the former was organized at a relatively late date. Since they are here so different it is not a priori certain that both combine equally to produce the fact we are studying. Anyway, if both contribute to it this cannot be in the same manner, nor probably in the same measure. Thus, we must investigate whether both take part and, if so, the share of each.

A proof of the slight effect of marriage is the fact that the marriage rate has changed very little since the first of the century, while suicide has tripled. From 1821 to 1830 there were 7.8 marriages annually per 1,000 inhabitants, 8 from 1831 to 1850, 7.9 in 1851-60, 7.8 from 1861 to 1870, 8 from 1871 to 1880. During this time the suicide-rate per million inhabitants rose from 54 to 180. From 1880 to 1888 the marriage rate declined slightly (7.4 instead of 8), but this decrease is unrelated to the enormous increase of suicides, which rose more than 16 percent from 1880 to 1887. Besides, during the period 1865-88, the average marriage rate of France (7.7) is almost the same as that of Denmark (7.8) and Italy (7.6); yet these countries are as different as possible from the point of view of suicide.

But we have a much more certain way of measuring exactly the real influence of conjugal association upon suicide; that of observing it when reduced to its own isolated strength, or in families without children.

During the years 1887-91, a million husbands without children accounted annually for 644 suicides. To know how much the marriage status, alone and without reference to the family, insures against suicide, one has only to compare this figure with that of the unmarried men of the same average age. This comparison Table XXI permits us to make, as not the least important of its information. The average age of married men was then as now 46 years, 8 and ⅓ months. A million unmarried men of this age have about 975 suicides. Now 644 is to 975 as 100 is to 150, that is, sterile husbands have a coefficient of preservation of only 1.5; they commit suicide only a third less often than unmarried of the same age. Quite otherwise when there are children. A million husbands with children annually show during this period only 336 suicides. This number is to 975 as 100 is to 290; that is, when the marriage produces children the coefficient of preservation is almost doubled (2.90 instead of 1.5).

Conjugal society therefore plays only a slight role in the immunity of married men. We have in the preceding calculation even made this role somewhat larger than it really is. We have assumed that childless husbands have the same average age as husbands in general, whereas they are certainly younger. For among their ranks are all the youngest husbands, who are without children not because they are hopelessly sterile, but because they have married too recently to have any. On the average, a man has his first child not before 34 years of age, and yet he marries at about 28 or 29 years of age. The part of the married population from 28 to 34 years of age is thus almost entirely in the category of the childless, which lowers the average age of these latter; therefore we must certainly have exaggerated in estimating it at 46. But in that case the unmarried men with whom they should have been compared are not those of 46 but younger, who consequently commit suicide less often than the others. So the coefficient of 1.5 must be a little too high; if we knew exactly the average age of childless husbands, their aptitude for suicide would surely approach that of unmarried men still more than the above figures indicate.

The limited influence of marriage is well shown, moreover, in that widowers with children are in a better situation than husbands without them. The former indeed, show 937 suicides per million. Now they are 61 years, 8 and ⅔ months on the average. The rate of unmarried men of the same age (see Table XXI) is between 1,434 and 1,768, or about 1,504. This number is to 937 as 160 is to 100. Widowers, when they have children, thus have a coefficient of preservation of at least 1.6, superior to that of childless husbands. Moreover, we have under rather than overestimated this figure. For widowers with children are certainly older than widowers in general. The latter, indeed, include all whose marriage was without issue only because of premature end by death, that is, the youngest. Widowers with children should therefore really be compared with unmarried men above 62 years (who, because of their age, have a stronger tendency to suicide). This comparison would clearly only emphasize their immunity.

To be sure, this coefficient of 1.6 is definitely below that of husbands with children, 2.9; the difference is not quite 45 per cent. Thus, matrimonial society by itself might be thought to have more effect than we have granted it, since at its conclusion the immunity of the husband surviving is so far reduced. But this loss is only in slight degree to be ascribed to the dissolution of marriage. Proof of this is that where there are no children widowhood produces far lesser effects. A million childless widowers show 1,258 suicides, a number related to 1,504, the contingent of sixty-two-year-old unmarried men, as 100 is to 119. Thus the coefficient of preservation is still about 1.2 which is little below that of husbands also childless, 1.5. The former of these figures is only 20 per cent less than the second. Accordingly, when a wife’s death has no other effect than to break the conjugal bond, it has no strong repercussion on the suicidal tendency of the widower. Marriage during its existence must therefore only slightly aid in restraining this tendency, since the latter shows no greater increase with the end of marriage.

The reason why widowhood is relatively more disastrous when the union has been fruitful must be sought in the existence of the children. Of course in a way the children attach the widower to life, but at the same time they make the crisis through which he is passing more intense. For not only is the conjugal relation destroyed; but precisely because a domestic society here exists, there is an impairment of its functioning too. An essential element is lacking and the whole machine is thrown out of gear. To reestablish the lost equilibrium the husband has to shoulder a double burden and perform functions for which he is unprepared. Thus he loses advantages which were his throughout the duration of the marriage. It is not because his marriage is ended but because the family which he heads is disorganized. The departure, not of the wife but of the mother, causes the disaster.

But the slight effect of marriage appears with special clarity in the woman’s case when it does not find its natural fulfillment in children. A million childless wives show 221 suicides; a million unmarried women of the same age (between 42 and 43 years) only 150. The first of these numbers is to the second as 100 is to 67; the coefficient of preservation thus falls below unity and equals .67, that is, it has really become a coefficient of aggravation. In France, then, married but childless women commit suicide half again as often as unmarried women of the same age. We have already noticed that in general the wife profits less from family life than the husband. Now we see the cause of this; in itself conjugal society is harmful to the woman and aggravates her tendency to suicide.

If most wives have, nevertheless, seemed to enjoy a coefficient of preservation, this is because childless households are the exception and consequently the presence of children in most cases corrects and reduces the evil effects of marriage. Even so these effects are only reduced. A million women having children show 79 suicides; comparing this figure with the one giving the suicide-rate of unmarried woman of 42 years of age as 150, the wife is found to benefit, even when she is also a mother, only by a coefficient of preservation 1.80, 35 per cent lower, therefore, than that of fathers. With respect to suicide, we must therefore disagree with the following proposition of Bertillon: “When woman enters the conjugal state she gains from the association more than man; but she necessarily suffers more than man when she leaves it.”

III

The immunity of married persons in general is thus due, wholly for one sex and largely for the other, to the influence not of conjugal society but of the family society. However, we have seen that even if there are no children, men at least are protected in the proportion of 1 to 1.5. A reduction of 50 suicides from 150, or 33 per cent, though considerably below that achieved when the family is complete, nevertheless is not a negligible quantity and its cause should be understood. Is it due to the special benefits bestowed by marriage on the male sex, or is it not rather a result of matrimonial selection? For although it has been shown that the latter does not play the dominant role attributed to it, it has not been proven to be wholly without influence.

One fact at first sight even seems to prove this hypothesis. We know that the coefficient of preservation of childless husbands partially survives marriage; it falls merely from 1.5 to 1.2. Now, this immunity of childless widowers evidently cannot be attributed to widowhood, which in itself does not tend to reduce the proclivity to suicide but on the contrary to confirm it. It thus results from an anterior cause, though this seems unlikely to be marraige, since it continues to act even when marriage is dissolved by the wife’s death. May it not then consist in some inherent quality of the husband which conjugal selection makes prominent but does not create? As it existed before marriage and is independent of it, it might well outlast the latter. If the population of husbands is an elite, the same must be true of widowers. To be sure, this congenital superiority has less effect upon the latter, since they are less protected against suicide. But the shock of widowhood may be considered as partially neutralizing this preventive influence and blocking its full results.

But for this explanation to be acceptable, it must be applicable to both sexes. Some trace at least of this natural predisposition should accordingly be found among married women, which, other things being equal, would preserve them from suicide more than the unmarried. Now the very fact that they commit suicide, if childless, more than unmarried women of the same age, is opposed to the hypothesis that they are endowed from birth with a personal coefficient of preservation. One might, however, grant that this coefficient exists for women as well as for men, but that it is wholly annulled during marriage by the unfortunate effect of marriage on the wife’s moral constitution. But if its effects were only restrained and concealed by the sort of moral decline of women on entering into conjugal society, they should reappear on the dissolution of this society, or in widowhood. Freed from the depressing influence of the matrimonial yoke, women should then recover all their advantages and finally assert their inherent superiority to those of their sisters who have not achieved marriage. In comparison with unmarried women, in other words, the childless widow should have a coefficient of preservation at least approaching that of the childless widower. This is not so. A million childless widows show annually 322 suicides; a million unmarried women of 60 (the average age of widows) show only between 189 and 204, or about 196. The first is to the second number as 100 to 60. Widows without children thus have a coefficient below unity, or a coefficient of aggravation; it is 0.60, slightly lower even than that of childless wives (0.67). It is therefore not marriage which prevents childless wives from showing the natural indisposition to suicide attributed to them.

Perhaps it will be objected that the obstacle to the complete reestablishment of the fortunate qualities whose expression is interrupted by marriage, is that widowhood is, for women, an even worse status. Indeed the idea is widespread that a widow is in a more critical position than a widower. The moral and economic difficulties are stressed which face her when she is compelled to provide all by herself, for her own existence as well as for the needs of an entire family. This opinion has even been considered proved by facts. According to Morselli, statistics prove that woman, during widowhood, is closer to man in her aptitude for suicide than she is during marriage; and since, when married, she is already nearer the male sex in this respect than when unmarried, therefore widowhood results in placing woman in the most disadvantageous position. Supporting this thesis, Morselli cites the following figures relating only to France, but to be found among all European peoples with slight variations:

Woman’s share in the suicides committed by both sexes in the state of widowhood seems in fact much greater than in the suicides of married persons. Does not this prove that widowhood is much more difficult for women than marriage? If so, it is not astonishing that, once a widow, the good effects of her qualities are even more prevented from appearing than before.

Unfortunately, this supposed law is based on an error of fact. Morselli has forgotten that there are everywhere twice as many widows as widowers. In France, there are in round numbers two million of the former to only one million of the latter. In Prussia, according to the census of 1890, 450,000 widowers are found and 1,319,000 widows; in Italy, 571,000 and 1,322,000 respectively. Under these conditions the share of widows is naturally higher than that of wives who are obviously of the same number as husbands. To obtain information from the comparison the two populations should be set up as equal. But if this precaution is taken, results contrary to those of Morselli are obtained. At the average age of the widowed or 60 years, a million wives show 154 suicides and a million husbands 577. Woman’s share is therefore 21 per cent. It diminishes perceptibly in widowhood. Indeed, a million widows show 210 cases, a million widowers 1,017; whence it follows that of 100 suicides of widowed persons of both sexes women contribute only 17. The share of men on the contrary rises from 79 to 83 per cent. Man thus loses more than woman in passing from marriage to widowhood, since he does not preserve certain of the advantages which he owed to the conjugal state. There is thus no reason to assume that this change of situation is less trying or disturbing for him than for her; the opposite is the case. Besides, we know that the mortality of widowers far exceeds that of widows; and the same is true of their marriage rates. That of widowers is at every age three or four times as great as the race of unmarried men, while that of widows is only slightly above that of unmarried women. Women are therefore as reluctant to face second marriage as men are eager. It would not be so if the state of widowhood sat lightly upon men and if women, on the contrary, had in that state to face as many difficulties as has been said to be the case.

But if nothing in widowhood particularly paralyzes woman’s natural advantages that pertain to her solely as matrimonially elect, and if these advantages do not manifest themselves by any definite sign, there is no reason for assuming their existence. The hypothesis of matrimonial selection is therefore wholly inapplicable to the female sex. Nothing justifies the supposition that a woman entering marriage has a constitutional advantage which preserves her to a certain degree from suicide. Consequently the same supposition is just as unfounded for men. The coefficient of 1.5 of childless husbands does not result from their belonging to the healthiest portion of the population; it can only be an effect of marriage. Conjugal society, so disadvantageous for women, must, even in the absence of children, be admitted to be advantageous for men. Those who enter it are not an aristocracy of birth; they do not bring to marriage, as an existing quality, a temperament disinclining them to suicide, but acquire it by living the conjugal life. At least, if they have some natural advantages these can be only very vague and indeterminate; for they are without influence until the advent of certain other conditions. So true is it that suicide does not principally depend upon the congenital qualities of individuals but upon causes exterior to and dominating them!

There is, however, a final difficulty to be solved. If this coefficient of 1.5, independent of the family, is due to marriage, how does it survive marriage and reappear at least in attenuated form (1.2) in the childless widower? If the theory of matrimonial selection which accounted for this survival is rejected, with what shall it be replaced?

It is sufficient to assume that the habits, tastes, and tendencies formed during marriage do not disappear on its dissolution; and nothing is more plausible than this hypothesis. If the married man, then, even if childless, feels a relative security from suicide, he must inevitably preserve some of this feeling when a widower. Only, as widowhood does involve a certain moral shock and since, as we shall see later, any loss of equilibrium inclines to suicide, this disposition, though remaining, is weakened. Inversely, but for the same reason, since a childless wife more often commits suicide than if she had remained unmarried, once become a widow she retains this stronger propensity for suicide, even slightly reenforced by the distress and loss of equilibrium always accompanying widowhood. But, since the ill effects that marriage had upon her make this change of status more acceptable, the aggravation is very slight. The coefficient is lowered by only a few per cent (0.60 instead of 0.67).

This explanation is confirmed by the fact that it is only a particular instance of a more general proposition which may be formulated thus: In an identical society, the tendency to suicide in the state of widowhood is for each sex a function of the suicidal tendency of the same sex in the state of marriage. If the husband is highly protected, the widower is too, although of course to a lesser degree; if the former is only slightly protected from suicide, the latter is not thus protected at all or only very little. To assure ourselves of the accuracy of this proposition we need only refer to Tables XX and XXI and the conclusions drawn from them. We there found that one sex is always more favored than the other in both marriage and widowhood. Now, the one more privileged in the first of these conditions preserves its privilege in the second. In France, husbands have a higher coefficient of preservation than wives; that of widowers is similarly higher than that of widows. In Oldenburg the opposite is true among married couples: the wife has a higher immunity than the husband. The same inversion occurs between widowers and widows.

But as these two single cases might with some justice be considered an insufficient proof and as, on the other hand, statistical publications do not give us the necessary data to verify our proposition in other countries, we have resorted to the following procedure to extend the scope of our comparisons: we have calculated separately the suiciderate for each age-group and marital status in the department of the Seine on the one hand, and on the other in all the rest of the departments combined. The two social groups, thus isolated from each other, are sufficiently different for us to expect their comparison to be instructive. And family life actually does have very different effects upon suicide in them (see Table XXII). In the departments the husband has much more immunity than the wife. In only four age-groups does the former’s coefficient descend below 3, while the wife’s never reaches 2; the average is in one case 2.88, in the other 1.49. In the Seine the reverse is true; for husbands the coefficient averages only 1.56 while it is 1.79 for wives. The very same inversion is found between widowers and widows. In the provinces the average coefficient of widowers is high (1.45), that of widows much lower (0.78). In the Seine, on the contrary, the second is higher, rising to 0.93, close to unity, while the other falls to 0.75. Thus, whichever the favored sex, widowhood regularly corresponds to marriage.

More than this, if the key is sought to the variation of the coefficient of husbands from one social group to another and if the same study is then made for widowers, the following surprising results are obtained:

Husbands’ coefficient in provinces = 2.88 = 1.84 Husbands’ coefficient in the Seine 1.56

Widowers’ coefficient in provinces = 1.45 = 1.93 Widowers’ coefficient in the Seine = 0.75

and for women:

Wives’ coefficient in the Seine = 1.79 = 1.20 Wives’ coefficient in provinces 1.49

Widows’ coefficient in the Seine = 0.93 = 1.19 Widows’ coefficient in provinces 0.78

The numerical proportions are for each sex pretty nearly equal; for women, the equality, in fact, is almost absolute. Thus, not only does the coefficient of widowers follow suit when that of husbands rises or sinks, but it even increases or decreases in exactly the same measure. These relations may be expressed in a form still more clearly confirmative of the law we have stated. They imply, in fact, that everywhere, whichever the sex, widowhood decreases the immunity of the surviving partner in a constant proportion:

Husbands in provinces = 2.88 =1.98 Widowers in provinces 1.45 Husbands in Seine = 1.56 = 2.0

Widowers in Seine 0.75 Wives in provinces = 1.49 = 1.91 Widows in provinces 0.78 Wives in Seine = 1.79 = 1.92 Widows in Seine 0.93

The coefficient of widowed persons is about half that of married persons. It is thus no exaggeration to say that the aptitude for suicide of widowed persons is a function of the corresponding aptitude of married persons; in other words, the former is in part a consequence of the latter. But since marriage adds to the husband’s immunity, even without children, it is not surprising that the widower should retain a portion of this fortunate disposition.

At the same time that it solves the question we had asked ourselves, this result casts some light on the nature of widowhood. In fact, it teaches us that widowhood in itself is not a hopelessly disadvantageous condition. It is very often better than bachelorhood. To be truthful, the moral constitution of widowers and of widows is not at all specific, but depends on that of married people of the same sex and in the same country. It is only a prolongation of this. If you will tell me how marriage and family life in a given society affect men and women, I will tell you what widowhood does for each. Although the crisis of widowhood is more grievous where marriage and domestic society are both felicitous, by fortunate compensation people are better equipped to face it; and, inversely, this crisis is less grave where the matrimonial and family constitution leave more to be desired, but in return people are less equipped to resist it. Thus, in societies where man benefits more from the family than woman, he suffers more when left alone but is at the same time better able to endure it, because the salutary influences which he has undergone have made him more averse to desperate resolutions.

IV

The following table summarizes the facts just established:

From this table and the preceding remarks it appears that marriage has indeed a preservative effect of its own against suicide. But it is very limited and also benefits one sex only. Useful as it has been to attest its existence—and this usefulness will be better understood in a later chapter—the fact remains that the family is the essential factor in the immunity of married persons, that is, the family as the whole group of parents and children. Of course, since husband and wife are members, they too share in producing this result, however not as husband or wife but as father or mother, as functionaries of the family association. If the disappearance of one increases the chances that the other may commit suicide, it is not because the bonds uniting them personally are broken, but because a family disaster occurs, the shock of which the survivor undergoes. Reserving the special effect of marriage for later study, we shall say that domestic society, like religious society is a powerful counteragent against suicide.

This immunity even increases with the density of the family, that is with the increase in the number of its elements.

This proposition we have already stated and proved in an article appearing in the Revue philosophique of November 1888. But the insufficiency of statistical data then at our disposal did not permit as strict a proof as was desirable. We did not know the average number in family establishments either throughout France or in each department. Thus we had to assume that family density depended solely on the number of children and—this number itself not being indicated in the census—we had to estimate it indirectly by employing what demography terms the physiological increase, or the annual excess of births over a thousand deaths. To be sure, this substitution was not unreasonable, for where the increase is high, families in general can hardly be other than dense. However, this consequence is not inescapable and often does not occur. Where children habitually leave their parents early, either to emigrate, or to settle elsewhere, or for other reasons, the family density has no reference to their number. In fact, the home may be deserted no matter how fruitful the marriage has been. This happens both in cultured surroundings where the child is early sent away from home to commence or complete his education, and in impoverished neighborhoods where premature dispersion is necessitated by the hardships of existence. On the other hand, the family may include a moderate or even a large number of elements in spite of a merely average birth rate if the unmarried adults or even the married children continue to live with their parents and form a single domestic society. For all these reasons no exact measure is possible of the relative density of family groups without knowledge of their actual composition.

The census of 1886, the results of which were not published until the end of 1888, gave us this knowledge. If we study from its data the relation in the different French departments between suicide and the actual average of family members, the following are the results:

As suicides diminish, family density regularly increases.

Instead of comparing averages, if we analyze the content of each group we shall find only confirmation of this conclusion. In fact, for all France, the average membership is 39 persons per 10 families. If then we ask how many departments there are above or below the average in each of these 6 classes, we shall find them to be composed as follows:

The group with most suicides includes only departments with family numbers below the average. Gradually and most regularly, the relation is reversed until the inversion is complete. In the last class, where suicides are few, all the departments have a family density above average.

The two maps (Appendices) also have the same general configuration. The region where families have least density definitely has the same limits as that of most frequent suicides. It also occupies the North and East and extends, on one side, to Brittany, on the other to the Loire. In the West and South, on the contrary, where there are few suicides, the family generally has large numbers. This relation even recurs in certain details. In the northern region, two departments are notable for their low aptitude for suicide, the Nord and Pas-de-Calais, a fact so much the more surprising as the Nord is highly industrial and intense industrialization favors suicide. The same peculiarity appears on the other map. In these two departments family density is high, though very low in all neighboring departments. In the South, we find on both maps the same dark area formed by the Bouches-du-Rhône, the Var, Alpes-Maritimes, and, to the West, the same light area formed by Brittany. The irregularities are exceptional and never prominent; considering the great number of factors which can affect so complex a phenomenon, such a general agreement is significant.

The same inverse relation reappears in the way both phenomena have evolved in time. Suicide has constantly increased since 1826 and the birth-rate has decreased. From 1821 to 1830 the latter was still 308 births for 10,000 inhabitants; during the period 1881-88 it was only 240, and the decrease was uninterrupted in the interval. At the same time, there appears a tendency for the family to break up and disperse more and more. From 1856 to 1886, the number of households increased by 2 millions in round figures; regularly and steadily it rose from 8,796,276 to 10,662,423. Yet, during the same time-interval, the population increased only by two million persons. Each family therefore includes a smaller membership.

Facts thus are far from confirming the current idea that suicide is due especially to life’s burdens, since, on the contrary, it diminishes as these burdens increase. This is a consequence of Malthusianism not foreseen by its author. When he urged control of the numbers in families, he felt that this restriction was at least in some cases necessary to general well-being. Actually, it is so much a source of the reverse condition that it diminishes the human desire to live. Far from dense families being a sort of unnecessary luxury appropriate only to the rich, they are actually an indispensable staff of daily life. However poor one is, and even solely from the point of view of personal interest, it is the worst of investments to substitute wealth for a portion of one’s offspring.

This result agrees with the one we had reached before. Why does family density have this effect upon suicide? In reply one could not refer to the organic factor; for though absolute sterility has primarily physiological causes, insufficient fecundity has not, being usually voluntary and depending on a certain state of mind. Family density, moreover, measured as we have measured it, does not depend exclusively on the birth-rate; we have seen that where there are few children other elements may take their place and, vice versa, that their number may be of no significance if they do not actually and consistently share in the group life. Nor should this preservative virtue be ascribed to the special feelings of parents for their immediate descendants. Indeed, to be effective these very feelings presuppose a certain state of domestic society. They cannot be powerful if the family has broken up. It is therefore because the functioning of the family varies with its greater or less density, that the number of its component elements affects the suicidal tendency.

That is, the density of a group cannot sink without its vitality diminishing. Where collective sentiments are strong, it is because the force with which they affect each individual conscience is echoed in all the others, and reciprocally. The intensity they attain therefore depends on the number of consciences which react to them in common. For the same reason, the larger a crowd, the more capable of violence the passions vented by it. Consequently, in a family of small numbers, common sentiments and memories cannot be very intense; for there are not enough consciences in which they can be represented and reenforced by sharing them. No such powerful traditions can be formed there as unite the members of a single group, even surviving it and attaching successive generations to one another. Small families are also inevitably short-lived; and without duration no society can be stable. Not only are collective states weak in such a group, but they cannot be numerous; for their number depends on the active interchange of views and impressions, on the circulation of these views and impressions from one person to another; and, on the other hand, this very exchange is the more rapid the more persons there are participating in it. In a sufficiently dense society, this circulation is uninterrupted; for some social units are always in contact, whereas if there are few their relations can only be intermittent and there will be moments when the common life is suspended. Likewise, when the family is small, few relatives are ever together; so that domestic life languishes and the home is occasionally deserted.

But for a group to be said to have less common life than another means that it is less powerfully integrated; for the state of integration of a social aggregate can only reflect the intensity of the collective life circulating in it. It is more unified and powerful the more active and constant is the intercourse among its members. Our previous conclusion may thus be completed to read: just as the family is a powerful safeguard against suicide, so the more strongly it is constituted the greater its protection.

V

If statistics had not developed so late, it would be easy to show by the same method that this law applies to political societies. History indeed teaches us that suicide, generally rare in young societies in process of evolution and concentration, increases as they disintegrate. In Greece and Rome it makes its appearance with the overthrow of the old city-state organization and its progress marks successive stages of decadence. The same is observed in the Ottoman Empire. In France, on the eve of the Revolution, the turmoil which shook society with the disintegration of the older social system took shape in a sudden rush of suicides mentioned by contemporary authors.

But beside these historical data, suicide statistics, though hardly existing for longer than the past seventy years, supply us with some proofs of this proposition which are more precise than those given above.

Great political upheavals are sometimes said to increase the number of suicides. But Morselli has conclusively shown that facts contradict this view. All the revolutions which have occurred in France during this century reduced the number of suicides at the moment of their occurrence. In 1830, the total fell to 1,756 from 1,904 in 1829, amounting to a sudden drop of nearly 10 per cent. In 1848 the drop is no less; the annual figure changes from 3,647 to 3,301. Then, during the years 1848-49, the crisis which has just shaken France spreads through Europe; everywhere suicides decrease, and this decrease is more and more perceptible the more serious and prolonged the crisis. This appears in the following table:

In Germany public feeling ran much higher than in Denmark and the struggle lasted longer even than in France, where a new government was immediately formed; accordingly, the decrease is prolonged in the German states up to 1849. For that year, the decrease is 13 per cent in Bavaria, 18 per cent in Prussia; in Saxony, in a single year from 1848 to 1849, it is likewise 18 per cent.

In 1851, the same phenomenon does not occur in France, nor does it occur in 1852. Suicides remain stationary. But in France the coup d’etat of Louis Bonaparte has the usual effect; although it took place in December, the number of suicides fell from 483 in 1851 to 446 in 1852 (8 per cent), and even in 1853 they were 463. This fact would seem to prove that this governmental revolution disturbed Paris much more than the provinces, where it seems to have had little effect. Besides, generally speaking, the influence of such crises is always more noticeable in the capital than in the departments. In 1830, the decrease in Paris was 13 per cent (269 cases instead of 307 the year before and 359 the year after); in 1848, 32 per cent (481 cases instead of 698).

Mild as they are, mere election crises sometimes have the same result. Thus, in France the suicide record clearly shows the mark of the parliamentary crisis of May 16, 1877 and the resulting popular agitation, as well as of the 1889 elections which ended the Boulanger agitation. In proof, we need only compare the monthly distribution of suicides in these two years with that of the years immediately before and after.

During the first months of 1877 suicides were more numerous than in 1876 (1,945 cases from January to April instead of 1,784) and the rise continues in May and June. Only at the end of the last-named month are the Chambers dissolved, the electoral period actually if not legally begun; this is probably the moment when political passions were most excited, for they were bound subsequently to subside somewhat due to time and weariness. Accordingly, in July, instead of continuing to surpass those of the preceding year, suicides are 14 per cent below them. Except for a slight pause in August, the drop continues to October, though less strongly. The crisis is ending. Immediately upon its conclusion, the rise, momentarily interrupted, is resumed. In 1889 the phenomenon is yet more pronounced. The Chamber is dissolved at the beginning of August; the excitement of the election period begins at once and lasts to the end of September, the time of the elections. An abrupt decrease of 12 per cent, compared with the corresponding month of 1888, occurs in August and lasts into September but stops abruptly in October when the struggle is ended.

Great national wars have the same effect as political disturbances. In 1866 war breaks out between Austria and Italy, and suicides drop 14 per cent in both countries.

In 1864 it was the turn of Denmark and Saxony. In the latter state suicides, which numbered 643 in 1863, fell to 545 in 1864 (16 per cent), only to return to 619 in 1865. As to Denmark, since we do not know the number of suicides for 1863, we cannot compare that of 1864 with it; but we do know that the figure for the second year (411) is the lowest since 1852. And as there is a rise to 451 in 1865, this figure of 411 very probably betokens a considerable drop.   a

The war of 1870-1871 had the same results in France and Germany:

This decrease might perhaps be considered due to the drafting of a part of the civilian population in war-time and the fact that it is very hard to keep track of suicides in an army in the field. But women as well as men contribute to this decrease. In Italy, suicides of women drop from 130 in 1864 to 117 in 1866; in Saxony, from 133 in 1863 to 120 in 1864 and 114 in 1865 (15 per cent). In the same country there is a no less considerable drop in 1870; from 130 in 1869 suicides fall to 114 in 1870 and remain at the same level in 1871; the decrease is 13 per cent, more than that of contemporary suicides of men. While 616 women had killed themselves in Prussia in 1869, there were only 540 such suicides in 1871 (13 per cent). It is common knowledge, besides, that young men capable of bearing arms furnish only a small contingent of suicides. Only six months of 1870 were occupied by the war; at this period, in time of peace, a million Frenchmen of from 25 to 30 years of age would have showed at most about 100 suicides, whereas the reduction between 1870 and 1869 is 1,057 cases.

The question has also been raised whether the cause of this momentary drop at a time of crisis might not be that the record of suicides was less exactly kept because of the paralysis of administrative authority. Numerous facts, however, show that this accidental cause does not adequately explain the matter. First, the widespread occurrence of the phenomenon. It appears among conquerors as well as vanquished, invaders and invaded alike. Furthermore, when the shock was very violent, its effects persisted for a considerable time after the event. Suicides increase slowly; some years pass before their return to their point of departure; this is true even in countries where in normal times they increase with annual regularity. Though partial omissions are of course possible and even likely at such times of trouble, the drop revealed by the statistics is too steady to be attributed to a brief inadvertence of administration as its principal cause.

But the best proof that we confront a phenomenon of social psychology and not a mistake in accounting, is that not all political or national crises have this influence. Only those do which excite the passions. We have already noted that revolutions in France have always had more affect on suicide in Paris than in the departments; yet the administrative upheaval was the same in the provinces and in the capital. But this sort of event always has much less interest for the provincial than for the Parisian, its author and participant from a closer vantage point. Likewise, while great national wars such as that of 1870-71 have had a strong influence on the current of suicide in both France and Germany, purely dynastic wars such as the Crimean or Italian, which have not violently moved the masses, have had no appreciable effect. There even occurred a considerable rise in 1854 (3700 cases against 3,415 in 1853). The same fact is observed in Prussia at the time of the wars of 1864 and 1866. The figures are stationary in 1864 and rise slightly in 1866. These wars were due wholly to the initiative of politicians and had not aroused public feeling like that of 1870.

From this point of view it is interesting to note that in Bavaria the year 1870 did not have the same effects as in the other countries of Germany, especially North Germany. More suicides were recorded in 1870 in Bavaria than in 1869 (452 against 425). Only in 1871 is there a slight decrease; it continues somewhat in 1872 when there are only 412 cases, which, however, entails a lowering of only 9 per cent by comparison with 1869 and 4 per cent with 1870. Yet Bavaria took the same important part as Prussia in military events; it, too, mobilized its whole army and the administrative disturbance must have been no less. It simply did not take the same moral share in events. Actually it is well known that Catholic Bavaria is, of all Germany, the country which has always lived a life most its own and been most jealous of its autonomy. It shared in the war through the will of its king but without enthusiasm. Therefore it resisted the great social movement then agitating Germany much more than the other allies; and so the reaction was felt there only later and less strongly. Enthusiasm was delayed and inconsiderable. It required the breath of glory wafted over Germany on the morrow of the victory of 1870 to warm somewhat the hitherto cold and unresponsive land of Bavaria.

This fact may be compared with the following, of similar significance. In France during the years 1870-71, suicide diminished only in the cities:

Recordings of suicides must, however, have been more difficult in the country than in the city. The true reason for this difference accordingly lies elsewhere. The war produced its full moral effect only on the urban population, more sensitive, impressionable and also better informed on current events than the rural population.

These facts are therefore susceptible of only one interpretation; namely, that great social disturbances and great popular wars rouse collective sentiments, stimulate partisan spirit and patriotism, political and national faith, alike, and concentrating activity toward a single end, at least temporarily cause a stronger integration of society. The salutary influence which we have just shown to exist is due not to the crisis but to the struggles it occasions. As they force men to close ranks and confront the common danger, the individual thinks less of himself and more of the common cause. Besides, it is comprehensible that this integration may not be purely momentary but may sometimes outlive its immediate causes, especially when it is intense.

VI

We have thus successively set up the three following propositions:

Suicide varies inversely with the degree of integration of religious society.

Suicide varies inversely with the degree of integration of domestic society.

Suicide varies inversely with the degree of integration of political society.

This grouping shows that whereas these different societies have a moderating influence upon suicide, this is due not to special characteristics of each but to a characteristic common to all. Religion does not owe its efficacy to the special nature of religious sentiments, since domestic and political societies both produce the same effects when strongly integrated. This, moreover, we have already proved when studying directly the manner of action of different religions upon suicide. Inversely, it is not the specific nature of the domestic or political tie which can explain the immunity they confer, since religious society has the same advantage. The cause can only be found in a single quality possessed by all these social groups, though perhaps to varying degrees. The only quality satisfying this condition is that they are all strongly integrated social groups. So we reach the general conclusion: suicide varies inversely with the degree of integration of the social groups of which the individual forms a part.

But society cannot disintegrate without the individual simultaneously detaching himself from social life, without his own goals becoming preponderant over those of the community, in a word without his personality tending to surmount the collective personality. The more weakened the groups to which he belongs, the less he depends on them, the more he consequently depends only on himself and recognizes no other rules of conduct than what are founded on his private interests. If we agree to call this state egoism, in which the individual ego asserts itself to excess in the face of the social ego and at its expense, we may call egoistic the special type of suicide springing from excessive individualism.

But how can suicide have such an origin?

First of all, it can be said that, as collective force is one of the obstacles best calculated to restrain suicide, its weakening involves a development of suicide. When society is strongly integrated, it holds individuals under its control, considers them at its service and thus forbids them to dispose wilfully of themselves. Accordingly it opposes their evading their duties to it through death. But how could society impose its supremacy upon them when they refuse to accept this subordination as legitimate? It no longer then possesses the requisite authority to retain them in their duty if they wish to desert; and conscious of its own weakness, it even recognizes their right to do freely what it can no longer prevent. So far as they are the admitted masters of their destinies, it is their privilege to end their lives. They, on their part, have no reason to endure life’s sufferings patiently. For they cling to life more resolutely when belonging to a group they love, so as not to betray interests they put before their own. The bond that unites them with the common cause attaches them to life and the lofty goal they envisage prevents their feeling personal troubles so deeply. There is, in short, in a cohesive and animated society a constant interchange of ideas and feelings from all to each and each to all, something like a mutual moral support, which instead of throwing the individual on his own resources, leads him to share in the collective energy and supports his own when exhausted.

But these reasons are purely secondary. Excessive individualism not only results in favoring the action of suicidogenic causes, but it is itself such a cause. It not only frees man’s inclination to do away with himself from a protective obstacle, but creates this inclination out of whole cloth and thus gives birth to a special suicide which bears its mark. This must be clearly understood for this is what constitutes the special character of the type of suicide just distinguished and justifies the name we have given it. What is there then in individualism that explains this result?

It has been sometimes said that because of his psychological constitution, man cannot live without attachment to some object which transcends and survives him, and that the reason for this necessity is a need we must have not to perish entirely. Life is said to be intolerable unless some reason for existing is involved, some purpose justifying life’s trials. The individual alone is not a sufficient end for his activity. He is too little. He is not only hemmed in spatially; he is also strictly limited temporally. When, therefore, we have no other object than ourselves we cannot avoid the thought that our efforts will finally end in nothingness, since we ourselves disappear. But annihilation terrifies us. Under these conditions one would lose courage to live, that is, to act and struggle, since nothing will remain of our exertions. The state of egoism, in other words, is supposed to be contradictory to human nature and, consequently, too uncertain to have chances of permanence.

In this absolute formulation the proposition is vulnerable. If the thought of the end of our personality were really so hateful, we could consent to live only by blinding ourselves voluntarily as to life’s value. For if we may in a measure avoid the prospect of annihilation we cannot extirpate it; it is inevitable, whatever we do. We may push back the frontier for some generations, force our name to endure for some years or centuries longer than our body; a moment, too soon for most men, always comes when it will be nothing. For the groups we join in order to prolong our existence by their means are themselves mortal; they too must dissolve, carrying with them all our deposit of ourselves. Those are few whose memories are closely enough bound to the very history of humanity to be assured of living until its death. So, if we really thus thirsted after immortality, no such brief perspectives could ever appease us. Besides, what of us is it that lives? A word, a sound, an imperceptible trace, most often anonymous, therefore nothing comparable to the violence of our efforts or able to justify them to us. In actuality, though a child is naturally an egoist who feels not the slightest craving to survive himself, and the old man is very often a child in this and so many other respects, neither ceases to cling to life as much or more than the adult; indeed we have seen that suicide is very rare for the first fifteen years and tends to decrease at the other extreme of life. Such too is the case with animals, whose psychological constitution differs from that of men only in degree. It is therefore untrue that life is only possible by its possessing its rationale outside of itself.

Indeed, a whole range of functions concern only the individual; these are the ones indispensable for physical life. Since they are made for this purpose only, they are perfected by its attainment. In everything concerning them, therefore, man can act reasonably without thought of transcendental purposes. These functions serve by merely serving him. In so far as he has no other needs, he is therefore self-sufficient and can live happily with no other objective than living. This is not the case, however, with the civilized adult. He has many ideas, feelings and practices unrelated to organic needs. The roles of art, morality, religion, political faith, science itself are not to repair organic exhaustion nor to provide sound functioning of the organs. All this supra-physical life is built and expanded not because of the demands of the cosmic environment but because of the demands of the social environment. The influence of society is what has aroused in us the sentiments of sympathy and solidarity drawing us toward others; it is society which, fashioning us in its image, fills us with religious, political and moral beliefs that control our actions. To play our social role we have striven to extend our intelligence and it is still society that has supplied us with tools for this development by transmitting to us its trust fund of knowledge.

Through the very fact that these superior forms of human activity have a collective origin, they have a collective purpose. As they derive from society they have reference to it; rather they are society itself incarnated and individualized in each one of us. But for them to have a raison d’etre in our eyes, the purpose they envisage must be one not indifferent to us. We can cling to these forms of human activity only to the degree that we cling to society itself. Contrariwise, in the same measure as we feel detached from society we become detached from that life whose source and aim is society. For what purpose do these rules of morality, these precepts of law binding us to all sorts of sacrifices, these restrictive dogmas exist, if there is no being outside us whom they serve and in whom we participate? What is the purpose of science itself? If its only use is to increase our chances for survival, it does not deserve the trouble it entails. Instinct acquits itself better of this role; animals prove this. Why substitute for it a more hesitant and uncertain reflection? What is the end of suffering, above all? If the value of things can only be estimated by their relation to this positive evil for the individual, it is without reward and incomprehensible. This problem does not exist for the believer firm in his faith or the man strongly bound by ties of domestic or political society. Instinctively and unreflectively they ascribe all that they are and do, the one to his Church or his God, the living symbol of the Church, the other to his family, the third to his country or party. Even in their sufferings they see only a means of glorifying the group to which they belong and thus do homage to it. So, the Christian ultimately desires and seeks suffering to testify more fully to his contempt for the flesh and more fully resemble his divine model. But the more the believer doubts, that is, the less he feels himself a real participant in the religious faith to which he belongs, and from which he is freeing himself; the more the family and community become foreign to the individual, so much the more does he become a mystery to himself, unable to escape the exasperating and agonizing question: to what purpose?

If, in other words, as has often been said, man is double, that is because social man superimposes himself upon physical man. Social man necessarily presupposes a society which he expresses and serves. If this dissolves, if we no longer feel it in existence and action about and above us, whatever is social in us is deprived of all objective foundation. All that remains is an artificial combination of illusory images, a phantasmagoria vanishing at the least reflection; that is, nothing which can be a goal for our action. Yet this social man is the essence of civilized man; he is the masterpiece of existence. Thus we are bereft of reasons for existence; for the only life to which we could cling no longer corresponds to anything actual; the only existence still based upon reality no longer meets our needs. Because we have been initiated into a higher existence, the one which satisfies an animal or a child can satisfy us no more and the other itself fades and leaves us helpless. So there is nothing more for our efforts to lay hold of, and we feel them lose themselves in emptiness. In this sense it is true to say that our activity needs an object transcending it. We do not need it to maintain ourselves in the illusion of an impossible immortality; it is implicit in our moral constitution and cannot be even partially lost without this losing its raison d’etre in the same degree. No proof is needed that in such a state of confusion the least cause of discouragement may easily give birth to desperate resolutions. If life is not worth the trouble of being lived, everything becomes a pretext to rid ourselves of it.

But this is not all. This detachment occurs not only in single individuals. One of the constitutive elements of every national temperament consists of a certain way of estimating the value of existence. There is a collective as well as an individual humor inclining peoples to sadness or cheerfulness, making them see things in bright or sombre lights. In fact, only society can pass a collective opinion on the value of human life; for this the individual is incompetent. The latter knows nothing but himself and his own little horizon; thus his experience is too limited to serve as a basis for a general appraisal. He may indeed consider his own life to be aimless; he can say nothing applicable to others. On the contrary, without sophistry, society may generalize its own feeling as to itself, its state of health or lack of health. For individuals share too deeply in the life of society for it to be diseased without their suffering infection. What it suffers they necessarily suffer. Because it is the whole, its ills are communicated to its parts. Hence it cannot disintegrate without awareness that the regular conditions of general existence are equally disturbed. Because society is the end on which our better selves depend, it cannot feel us escaping it without a simultaneous realization that our activity is purposeless. Since we are its handiwork, society cannot be conscious of its own decadence without the feeling that henceforth this work is of no value. Thence are formed currents of depression and disillusionment emanating from no particular individual but expressing society’s state of disintegration. They reflect the relaxation of social bonds, a sort of collective asthenia, or social malaise, just as individual sadness, when chronic, in its way reflects the poor organic state of the individual. Then metaphysical and religious systems spring up which, by reducing these obscure sentiments to formulae, attempt to prove to men the senselessness of life and that it is self-deception to believe that it has purpose. Then new moralities originate which, by elevating facts to ethics, commend suicide or at least tend in that direction by suggesting a minimal existence. On their appearance they seem to have been created out of whole cloth by their makers who are sometimes blamed for the pessimism of their doctrines. In reality they are an effect rather than a cause; they merely symbolize in abstract language and systematic form the physiological distress of the body social. As these currents are collective, they have, by virtue of their origin, an authority which they impose upon the individual and they drive him more vigorously on the way to which he is already inclined by the state of moral distress directly aroused in him by the disintegration of society. Thus, at the very moment that, with excessive zeal, he frees himself from the social environment, he still submits to its influence. However individualized a man may be, there is always something collective remaining—the very depression and melancholy resulting from this same exaggerated individualism. He effects communion through sadness when he no longer has anything else with which to achieve it.

Hence this type of suicide well deserves the name we have given it. Egoism is not merely a contributing factor in it; it is its generating cause. In this case the bond attaching man to life relaxes because that attaching him to society is itself slack. The incidents of private life which seem the direct inspiration of suicide and are considered its determining causes are in reality only incidental causes. The individual yields to the slightest shock of circumstance because the state of society has made him a ready prey to suicide.

Several facts confirm this explanation. Suicide is known to be rare among children and to diminish among the aged at the last confines of life; physical man, in both, tends to become the whole of man. Society is still lacking in the former, for it has not had the time to form him in its image; it begins to retreat from the latter or, what amounts to the same thing, he retreats from it. Thus both are more self-sufficient. Feeling a lesser need for self-completion through something not themselves, they are also less exposed to feel the lack of what is necessary for living. The immunity of an animal has the same causes. We shall likewise see in the next chapter that, though lower societies practice a form of suicide of their own, the one we have just discussed is almost unknown to them. Since their social life is very simple, the social inclinations of individuals are simple also and thus they need little for satisfaction. They readily find external objectives to which they become attached. If he can carry with him his gods and his family, primitive man, everywhere that he goes, has all that his social nature demands.

This is also why woman can endure life in isolation more easily than man. When a widow is seen to endure her condition much better than a widower and desires marriage less passionately, one is led to consider this ease in dispensing with the family a mark of superiority; it is said that woman’s affective faculties, being very intense, are easily employed outside the domestic circle, while her devotion is indispensable to man to help him endure life. Actually, if this is her privilege it is because her sensibility is rudimentary rather than highly developed. As she lives outside of community existence more than man, she is less penetrated by it; society is less necessary to her because she is less impregnated with sociability. She has few needs in this direction and satisfies them easily. With a few devotional practices and some animals to care for, the old unmarried woman’s life is full. If she remains faithfully attached to religious traditions and thus finds ready protection against suicide, it is because these very simple social forms satisfy all her needs. Man, on the contrary, is hard beset in this respect. As his thought and activity develop, they increasingly overflow these antiquated forms. But then he needs others. Because he is a more complex social being, he can maintain his equilibrium only by finding more points of support outside himself, and it is because his moral balance depends on a larger number of conditions that it is more easily disturbed.

Durkheim’s figure of 132 appears to be a misprint. The figure works out to 139.—Ed.

See Wagner. Die Gesetzmässigkeit, etc., p. 177.

See article, Mariage, in Dictionnaire encyclopédique des sciences medicales, 2nd series. See p. 50 ff.—On this question cf. J. Bertillon, Jr., Les célibataires, les veufs et les divorcés au point de vue du mariage in Revue scientifique, February, 1879.—Also an article in the Bulletin de la société d’ anthropologie, 1880, p. 280 ff.—Durkheim, Suicide et natalité, in Revue philosophique, November 1888.

We assume that the average age of these groups is the same as in France. The error which may result from this assumption is very slight.

If the two sexes are considered combined. The importance of this remark will appear below (Bk. II, Ch. 5, Par. 3).

See Bertillon, art., Mariage, in Dict. Encycl, 2d series, see p. 52. Morselli, p. 348.—Corre, Crime et suicide, p. 472.

Yet the labor of assembling these data, considerable if undertaken by an individual, might easily be accomplished by the official bureaus of statistics. All sorts of valueless information is given and only that omitted which, as will be seen below, might show the state of family life in the different European societies.

There are, to be sure, Swedish statistics reproduced in the Bulletin de démographie internationale, 1878, p. 195, giving these data. But they are useless. In the first place, widowed persons are there combined with unmarried persons, making the comparison relatively insignificant, for such different conditions must be distinguished. Moreover, we believe these statistics to be inexact. Here for example are some of their figures:

These figures have from the first seemed suspicious with regard to the tremendous degree of relative immunity enjoyed by married persons of advanced age, since they differ so from all facts known to us. To achieve verification that we deem indispensable, we have examined the absolute numbers of suicides committed by each age-group in Sweden at the same period. For men they are as follows:

Comparing these figures with the proportional numbers given above, the error committed becomes obvious. Actually, from 66 to 75 years, married and non-married persons show almost the same absolute number of suicides, whereas per 100,000 the former are supposed to kill themselves eleven times less often than the latter. For this to be true there would have to be at this age about ten times (exactly, 9.2 times) more married than non-married persons, that is, than widowed and unmarried combined. For the same reason, the married population above 75 should be exactly 10 times more numerous than the other. But that is impossible. At these advanced ages widowed persons are very numerous and, combined with unmarried persons, they are equal or even greater in number than married persons. This suggests what error has probably been committed. The suicides of unmarried and widowed persons must have been added together and the resulting total divided only by the figure for the unmarried population alone, while the suicides of married persons were divided by one for the widowed and married populations combined. What makes this probable is that the degree of immunity of married persons is extraordinary only at the advanced ages, or when the number of widowed persons becomes great enough seriously to falsify the resulting calculation. And the improbability is greatest after 75 years, or when widowed persons are very numerous.

See above p. 176.—To be sure, one might think that this unfavorable situation of married persons from 15 to 20 years is due to their average age being above that of unmarried persons in the same age-group. But what proves that there is a real aggravation is that the ratio of married persons of the following age-group (20 to 25 years) is five times less.

See Bertillon, art. Mariage, p. 43 ff.

There is a single exception; women of from 70 to 80 years, whose coefficient descends slightly below unity. The cause of this variation is the influence of the department of the Seine. In other departments (see Table XXII, p. 196), the coefficient of women of this age is above unity; but it must be noted that even in the provinces it is less than that of other ages.

Paris, 1888, p. 436.

J. Bertillon, Jr., article cited in the Revue scietifique.

To reject the hypothesis that the privileged position of married persons is due to matrimonial selection, the aggravation, which is supposed to result from widowhood, is sometimes mentioned. But we have just seen that no such aggravation exists by comparison with unmarried persons. Widowed persons kill themselves much less than non-married persons. Thus the argument does not carry.

These figures refer to France and the census of 1891.

We make this reservation, because this coefficient of 2.39 relates to the period from 15 to 20 years and because, since the suicides of wives are very rare at this age, the small number of cases which form the basis of these figures makes their exactness somewhat uncertain.

Usually, when the respective situation of the sexes in the two different sorts of marital status is thus compared, the effect of age is not carefully eliminated ; but this produces inexact results. Following the usual method, one would find in 1887-91, 21 suicides of married women to 79 of married men and 19 suicides of unmarried women to 100 of unmarried persons of all ages. These figures would give a false impression of the situation. The above table shows that the difference between the share of the married woman as against the unmarried woman is much greater at every age. Thus, for both the unmarried and married, the size of the difference between the sexes varies with the age-group, until in the age-group, 70 to 80, the size of the difference is about twice what it was at 20. Now, the unmarried population is almost wholly made up of persons below 30 years old. If, however, no account is taken of age, the difference resulting is actually that between unmarried men and women of about 30. But then, comparing this with the difference between married persons without respect to age, since the latter are on the average 50 years old, the comparison is really made with reference to married persons of this age. Thus the comparison is falsified and the error further aggravated by the fact that the difference between the sexes does not vary in the same way in both groups under the influence of age. Among the unmarried it increases more than among the married.

Durkheim fails to mention that in the 15-20 age-group for men the coefficient of preservation in Table XXI is 0.22.—Ed.

Thus one can also see from the preceding Table that the proportional share of wives in the suicides of married persons increasingly surpasses, with age, the share of unmarried women in the suicides of unmarried persons.

This statement by Durkheim must be carefully appraised in the light of Table XXI.—Ed.

Legoyt (op. cit., p. 175) and Corre (Crime et suicide, p. 475) nevertheless thought they could establish a relation between the variations of suicide and the marriage rate. Their error proceeds first from having considered too short a period, secondly from having compared the most recent years with an abnormal year, 1872, when the French marriage rate reached an unusual figure, unknown since 1813, because the gaps in the ranks of the married population caused by the war of 1870 had to be filled. No such reference can be a measure for the changes of the marriage rate. The same observation applies to Germany and to almost all the European countries. Something like an electric shock seems to have affected the marriage rate at the time. A great and abrupt rise is seen, prolonged occasionally as late as 1873, in Italy, Switzerland, Belgium, England, Holland. All Europe might be said to have been contributing to repair the losses of the two war-stricken countries. A tremendous fall naturally succeeded after some time which has not the significance ascribed to it (See Oettingen, Moralstatistik, supplement, Tables 1, 2 and 3).

See Levasseur, Population francaise, vol. II, p. 208.

According to the census of 1886, p. 123 of the Dénombrement.

See Annuaire statistique de la France, 15th vol., p. 43.

For the same reason, the age of husbands with children is above that of husbands in general and, consequently, the coefficient of preservation 2.9 should be considered somewhat below reality.

What Durkheim seems to mean here is that widowers with children compared to husbands with children are relatively worse off than widowers without children compared to husbands without children.—Ed.

similar difference exists between the coefficient of childless husbands and childless wives; it is much greater. The second (0.67) is 66 per cent lower than the first (1.5). The existence of children thus causes the wife to regain half the ground she loses by marriage. That is, if she benefits from marriage less than the man, she profits more than he from the family, that is, the children. She is more sensitive than he to their happy influence.

Article Mariage, Dict. Encycl., 2d series, vol. V, p. 36.

Op. cit., p. 342.

See Bertillon, Les celibataires, les veufs, etc., Rev. scient., 1879.

Morselli also mentions in support of his thesis that on the morrow of war the suicides of widows show a much greater rise than those of unmarried women or wives. But it is merely that then the population of widows increases disproportionately; thus it naturally produces more suicides and this rise naturally persists until the restoration of equilibrium and the return to their normal level of the different sorts of marital status.

When there are children, the lowering of coefficient incident to both sexes due to widowhood is almost the same. The coefficient of husbands with children is 2.9; it becomes 1.6. That of women in the same circumstances from 1.89 becomes 1.06. The diminution is 45 per cent for the former, 44 per cent for the latter. That is, as we have said, widowhood produces two different effects; it disturbs 1. the conjugal society, 2. the family society. The former disturbance is much less felt by the woman than by the man, just because she profits less from marriage. But the second is felt far more by her; for she often finds it harder to take the husband’s place in the direction of the family than he does to replace her in her domestic functions. When there are children, therefore, a sort of compensation occurs which makes the suicidal tendency of the two sexes vary, as a result of widowhood, in the same proportions. Thus it is especially when there are no children that a widowed woman partially recovers the ground lost in the state of marriage.

From Table XXII it appears that in Paris, as in the provinces, the coefficient of husbands below 20 years is below unity; that is, for them there is aggravation. This confirms the law formulated above.

Evidently, when the female sex is the one more favored by marriage, the disproportion between the sexes is much less than when the husband has the advantage; a new confirmation of a remark made above.

M. Bertillon (article cited in the Revue scientifique), had already given the suicide-rate for the different categories of marital status with and without children. He found the following results:

These figures refer to the years 1861-68. Given the general increase in suicides, they confirm our own figures. But as the lack of such a table as our Table XXI allowed no comparison of husbands and widowers with unmarried persons of the same age, no precise conclusion could be drawn as to the coefficients of preservation. We are also in doubt as to whether they refer to the entire country. Actually the French Bureau of Statistics assures us that the distinction between childless couples and those with children was never made in the census before 1886, except in 1855 for the departments exclusive of the Seine.

See Book II, Chap. V, 3.

See Dénombrement de 1886, p. 106.

The word “density” has just been used in a somewhat different sense from that usually given it in sociology. Generally, the density of a group is defined not as a function of the absolute number of associated individuals (which is rather called volume ) but of the number of individuals actually in reciprocal relationship in one and the same social volume. (See Durkheim, E., Regles de la Meth. social., p. 139). But in the case of the family the distinction between volume and density has no interest, since, due to the smallness of the group, all associated persons are in actual relationship.

Let us not confuse young societies capable of development with lower societies; in the latter, on the contrary, suicides are very frequent, as will appear in the following chapter.

Helvetius wrote in 1781: “Financial disorder and the change in the constitution of the state spread general fear. Numerous suicides in the capital give sad proof of this.” Quoted from Legoyt, p. 30. Mercier in his Tableau de Paris (1782) says that within 25 years the number of suicides tripled in Paris.

According to Legoyt, p. 252.

According to Masaryck, Der Selbstmord, p. 137.

Actually, the annual rate at this age in 1889-91 was only 396; the semi-annual rate about 200. From 1870 to 1890 the number of suicides at every age doubled.

Durkheim’s figures show a reduction from 5,114 in 1869 to 4,157 in 1870, which amounts to 957, not 1,057.—Ed.

Nor is it certain that this diminution of 1872 was caused by the events of 1870. The reduction of suicides scarcely made itself felt outside of Prussia beyond the actual period of hostilities. In Saxony the reduction of 1870, only 8 per cent, is not continued in 1871 and almost completely comes to an end in 1872. In the duchy of Baden it was confined to 1870; 1871, with its 244 cases, exceeds 1869 by 10 per cent. It thus seems that Prussia alone was seized with a sort of collective euphoria on the morrow of victory. The other states had less feeling for the increased glory and power resulting from the war, and social passions subsided with the end of the great national crisis.

See above, Book II, Ch. 2.

We say nothing of the ideal protraction of life involved in the belief in immortality of the soul, for (1) this cannot explain why the family or attachment to political society preserves us from suicide; and (2) it is not even this belief which forms religion’s prophylactic influence, as we have shown above.

This is why it is unjust to accuse these theorists of sadness of generalizing personal impressions. They are the echo of a general condition.