The choice of baby names can reveal how rapidly individuals adapt to new circumstances. A French genealogy database of 5.5 million individuals born between 1905 and 1925 permits an analysis of name transmission from father to son/daughter and uncle to nephew/niece as a reaction to the family disruptions caused by the exigencies of World War I. Immediately following mobilization, the rate of paternal name transmission increased by about 50 percent for both males and females, enhanced by a father’s level of risk as a soldier. Transmission of an uncle’s name increased following his death regardless of whether he died at war, though the phenomenon was significantly more sustained if he did so.
The structure of the family has long influenced the choice of baby names. “Naming for kin,” or the transmission of names from one generation to the next, is known to be more frequent when families have many children or when several generations cohabit. Naming for kin was the dominant practice in early modern Europe. From the late eighteenth century onward, however, it gradually declined, giving way to parental tastes and beliefs as baby name determinants. Circumstances when baby names are the outcome of parental choices enable the study of how people react to historical events. For instance, researchers have used names to investigate popular support for regime changes before the era of public-opinion polls, reactions to the Chinese Cultural Revolution, mass incarceration of U.S. citizens of Japanese descent following the attack on Pearl Harbor, and diplomatic crises. In all such cases, rapid changes in the popularity of specific names was evident. Witness the increased frequency of a president’s name in the United States following an assassination attempt on him. Less attention, however, has been paid to systematic modifications of naming behavior.1
World War I (WWI) was unquestionably a major social disruption in France. On August 2, 1914, France began mobilizing millions of young men for war, many among them to join fighting units. Couples were hastily separated. Most women immediately faced dire practical problems: The harvest was the most pressing issue in the countryside, and urban unemployment rose alarmingly. Violent combats began in the third week of August, and immediately thereafter, large parts of northeastern France underwent invasion by the German army. Since the seminal work of Becker, the attitude of European populations during the early phase of the conflict has been of special interest. Previous research has found that not until late July 1914 did the French people realize that war was imminent. They expected the fighting to be short-lived, finished by the end of the year. Furthermore, they were unaware of how much risk soldiers would face on the front lines before it became painfully obvious that modern weapons were causing unprecedented casualties.2
Diverse historical sources—administrative reports, letters, diaries, and newspapers—have all shed light on ordinary people’s perception of the war. None, however, are as systematic and direct as baby names, the exploration of which remains in its infancy in the context of WWI. Recently, a heightened popularity of a father’s name as a response to his death at the front was observed within an urban cohort of war orphans born between 1914 and 1916. Specifically, the prenatal loss of a father corresponded with a 2.7-fold increase in the probability of transmission of the father’s name, both to male and, in feminized form, female offspring, especially among officers’ dependents. The resulting hypothesis was that this naming practice was a concrete expression of a phenomenon that Mosse called the “cult of the fallen soldier,” which thus far had been the subject of case studies and is most visible in the war memorials (Monuments aux Morts) erected by municipalities after 1918. Previous work about orphans’ names was limited by a small sample size, an exclusively urban context, a lack of comparison to the antebellum period, and an exclusive focus on paternal names.3
This research note aims at a systematic description of attitudes during WWI by focusing on the determinants of kin (father and uncle) name transmission. Our objective is twofold—first, to document ordinary people’s perception of the severity of the conflict and, second, to investigate whether naming practices show any evidence that the cult of the fallen soldier was actually common.
Data and Methods
Geneanet—a crowd-sourced genealogy website where registered users can choose either to share their family tree (“open records”) or keep it private—granted us access to the data for its 5.5 million individuals born between 1905 and 1925 in France, as well as the attached information about their fathers and uncles: family name, first and middle names, sex, place and date of birth, place and date of death; for fathers: first and middle names, dates of birth and death; for uncles: line (maternal/paternal), first and middle names and date of death.4
Since individuals appear in the database whenever one of their descendants included them in his family tree, many records of the original data set are duplicates. To identify and merge such duplicates, we used the EM algorithm implemented by the RecordLinkage package of the R programming language. Due to the high computational cost of the de-duplication process, we stratified on sex and year of birth and ran it on a computer cluster. This procedure yielded N = 3,056,387 distinct individuals born in France between 1905 and 1925.5
We focused on first names. Hyphenated first names were considered as two separated names, the first of which we treated as the first name. In order to include females in the analysis, comparison of fathers’ and offspring’s names was limited to the first four letters. In other words, we considered transmission to have occurred when a father’s name was, say, Simon and a daughter’s name Simone.
Diachronic Analysis of Paternal Transmission
We studied the rate of transmission of paternal first names (the “transmission rate” for brevity) according to date of birth. We first focused on the beginning of the war, subdivided into three periods—the diplomatic crisis of July 1914 (the “July Crisis,” which started with the assassination of Archduke Franz Ferdinand on June 28); early mobilization (from the mobilization order on August 1 to the first significant fighting in the third week of August); and the combat phase (late August and September). We then analyzed the entire duration of the war. Excluding those for whom the exact date of birth or father’s name was missing, this analysis is based on N = 2,311,909 individuals.
Kin Name Transmission Following Kin Death
For the N = 1,421,152 individuals for whom father’s date of death was known, we studied the association between paternal-name transmission and age at father’s death, analyzing kin deaths in wartime and peacetime separately. Whereas a child’s birth cannot occur more than nine months after his/her father’s death, the same is evidently not true for uncles, who may have died many years before the birth of a nephew or niece. Thus, the investigation of naming after dead kin can encompass a much longer time scale. One methodological difficulty is that although a child has exactly one father, he or she potentially has many uncles, and conversely an uncle many nieces and nephews. Hence several definitions of the rate of transmission are possible. In this research note, we study the proportion of uncle–nephew/niece dyads for which the first four letters of the two names match, grouping the dyads according to the interval between uncle’s death and nephew/niece’s birth (“age at uncle’s death,” for short, even if avuncular death occurred prior to nephew/niece’s conception). The data set analyzed is that of uncle–nephew/niece dyads in which both uncle’s and child’s name are known (N = 6,055,890).
Paternal Transmission Rate During WWI
On the brink of WWI, about 12 percent of all baby boys and 4 percent of all baby girls received their father’s first name (Figure 1, top panel, inset). Transmission of father’s first name to boys was declining, as parents increasingly turned to fashion. By contrast, until the beginning of the 1920s, transmission of father’s first name to girls was becoming more common, partly because of a general increase in the popularity of male-name variants (for example, Marcelle or Paulette) that enabled fathers’ names to be transmitted as fashionable variants. No period before 1914—including the “Moroccan crises” of 1905 and 1911, when France and Germany threatened warfare—saw any disruption to these trends.
July Crisis and Early August 1914
We find no more evidence of an increase in the transmission rate during the July Crisis than we did for the Moroccan crises, not even for those born in the last week of July (Monday, July 27, to Sunday, August 2). By contrast, we observe a surge in the first week of August (Monday, August 3, to Sunday, August 9), immediately following the mobilization order but before any actual fighting: Whereas 14 percent (95 percent Confidence Interval [ci]: 11.9 percent–16 percent) of boys born during the last week of July took the name of their father, 17.7 percent (95 percent ci: 15.5 percent–19.9 percent) of boys born during the following week did. Female births underwent the same surge (Figure 1, lower panel).
First Combats and Subsequent Months
After the early August surge, the transmission rate remained stable until around May 1915, when it fell back to its prewar level. This phenomenon suggests hypotheses regarding both the mechanisms behind heightened name transmission and the sub-population involved: First, large-scale social change could not have been responsible for this phenomenon, since the social disruption that began in 1914 was hardly finished in May 1915; large parts of France were still under occupation and involved in combat. Second, the increased transmission that began in August 1914 mostly involved families with a father who had been mobilized. Children born after May 1915 were conceived after August 1914, hence by fathers who were still with their spouse when mobilization began; the majority of men available for conception of a child after August 1914 were those free from military obligation. Notably, for the same reason, the monthly proportion of orphans in the total population of births plummeted in May 1915 in a manner consistent with this “mobilization effect.”6
Heightened Transmission of Paternal First Name as a Measure of Risk Perception
Given the strong association of mobilization (and the subsequent separation of couples) in August 1914 with increased transmission, the question arises whether separation alone was the reason, or whether increased transmission was a function of the risks taken by the soldiers who were mobilized. The notion that the increase was correlated with father’s risk is consistent with our finding that for infants born to men older than thirty-five (and therefore expected to be mobilized in less exposed, “territorial,” units), the increase is negligible (Figure 2). Accordingly, the surge in transmission observed in August 1914 almost exclusively took place for babies born to fathers younger than thirty-five.
If a father’s level of risk was the driving factor behind higher transmission, name transmission and paternal death during the war must be correlated, since, by definition, fathers who died were statistically more at risk. We can assess name transmission among those children born to young fathers according to whether the father died during the war (excluding the trivial case of a father dying before the birth of the child; see below). Indeed, the transmission rate was 11 percent (95 percent ci: 10.8 percent–11.2 percent) for babies whose fathers died after the war, but 15.7 percent (95 percent ci: 14.9 percent–16.5 percent) for those whose father died during the war.
Naming for Dead Kin in War and Peace
Prenatal Loss of Father
Between 30 and 55 percent of all prenatal orphans (those still in the womb when their father died) received the first name of their father, regardless of gestational age at father’s death (Figure 3). Transmission rates were regularly, albeit moderately, higher for those whose father died during the war rather than in peacetime. Consistent with previous estimates, the loss of a father during his wife’s pregnancy leads in our sample to a 2.7-fold increase in the transmission rate. Due to a higher baseline value (transmission rate for a postnatal loss of father), the fold increase is lower in wartime (in boys, 2.2) than in peacetime (boys, 2.8). The loss of a father during the first trimester of his wife’s pregnancy further led to increased transmission, whereas a death occurring in the last month of her pregnancy was associated with a more modest, twofold increase. Competing hypotheses are that for these births, either (1) the mother was still not aware of the death of her husband, (2) the parents had already agreed upon a baby name, or (3) the need to link a baby to his or her father symbolically was less important when a pregnancy was already well-established at the time of the father’s death.
Prenatal Loss of Uncle
In our sample, the transmission rate of avuncular name to nephew was about 5 percent at baseline, that is, for uncles still alive at nephew’s birth (Figure 4). The transmission rate was distinctly higher in uncle–nephew/niece dyads in which the uncle died before the birth of his nephew/niece. The prenatal period also showed a marked difference between wartime and peacetime avuncular deaths. For deaths within the year preceding the birth, the rate of transmission to nephews was 14 percent (95 percent ci: 13.2 percent–14.8 percent) for a death in peacetime vs. 16 percent (95 percent ci: 15 percent–16.9 percent) for a death in wartime. The naming response to a wartime death of uncle was also more lasting than that to a peacetime death; the differences between wartime and peacetime avuncular deaths remain pronounced even as long as ten years after the uncle’s death.
The simple behavior of naming a child can yield considerable insight into the reactions to historical events. To investigate this behavior at the population level, we rely on digitized genealogies as the data sources. Due to their sample size and chronological coverage, such sources have become increasingly popular in fields such as population genetics and social history. Nonetheless, these sources have several shortcomings. The distribution of producers is skewed; a small number of users contributes a large number of records. Amateur genealogists tend to focus on the prestigious branch(es) of their family. Lower-class individuals are under-represented because of common-law marriages, unregistered births, illegitimate births, infant mortality, child abandonment, and everyday resistance to policing. If the precise size of the effects measured in our sample may not generalize to the population of all those born between 1905 and 1925 in France, we have no reason to believe that selectivity is the reason for our results, all based on large differences measured between groups within our data set.7
July 1914 retrospectively appears as the tragic month when European governments made the decisions that led to WWI. Though newspapers reported the assassination of Archduke Franz Ferdinand and the diplomatic tensions that followed, the press misunderstood and downplayed the importance of this stream of events for weeks. Unsurprisingly, no evidence for changes in paternal-transmission rates were in evidence during this period. The possibility of a war suddenly materialized during the last week of July. For instance, as late as July 24, 1914, the newspaper L’Humanité reported nothing about the international crisis until page 3, but on the very next day, Jean Jaurès, the head of the newspaper and a major socialist figure, wrote a front-page article announcing that Europe was on the brink of war, with only a “suprême chance de Paix” left. Nonetheless, even though unions and the Socialist party organized demonstrations in large cities that same week, we have no evidence for a consequential change in naming behavior at that point, suggesting French parents did not grasp the enormity of the threat. Not until actual mobilization did the naming pattern change. The study of paternal name transmission in July 1914 is largely congruent with previous work: As Becker stated, “When the church bells began to ring on 1 August and it became clear that this was not to warn of fire but to announce mobilisation for war, the first reaction was one of stunned shock and consternation.”8
Our analysis shows that suddenly, during the first weeks of August, before the first deadly military engagements, paternal names became more popular than ever. This development was too abrupt to have been driven by imitation; most likely, mothers were not even aware that a change in naming practices was underway. The change is quantitatively significant regarding children born to young fathers (most of whom were sent to fighting units) and negligible regarding children of older fathers (most of whom were sent to non-fighting units). A father’s level of risk, rather than separation of parents in itself, seems to have boosted father’s name transmission. The fact that the rate of transmission did not change in late August to early September is puzzling, since mothers who initially underestimated the risk of paternal death might well have re-estimated it once the first casualty rates became available. This stable degree of transmission supports the notion that parents correctly estimated a high risk of death from the outset of the war. Experience of the 1870 Franco-Prussian War or knowledge about the lethality of modern weapons gained from the Balkan wars may have been sufficient to inform people’s assessment of risk. At odds with previous narratives, we conclude that even in its earliest phase, ordinary people did not underestimate the severity of the conflict.
Localized family disturbance rather than large-scale social disruption seems to have been at the root of the change in naming practices. The war and the removal of young men from their village during mobilization, however, would seem to have overturned the established social hierarchies. By all rights, such a large-scale upheaval should have affected every parent of babies born after July 1914. Yet, after May 1915, the transmission rate fell back to its pre-war level, even as the war continued. The ongoing general disruption of the social order did not exert a greater influence on the parents. Furthermore, only a modest increase in paternal name transmission is evident for babies born to older fathers, mobilized in territorial units (“armée territoriale”). The change in naming behavior mainly pertained to families where the separation of parents entailed a heightened risk of death.
Regarding the sub-population of children who lost their father or an uncle, we could systematically investigate differences in mourning practices during war and peace. Many studies have already been devoted to the unprecedented mass grief induced by WWI, particularly in the postwar period when virtually every French town and village constructed war memorials, and many people made pilgrimages to military cemeteries. Documentation about the frequency and timing of intimate rituals has been more scattered. Anecdotal testimonies indicate that photographs and medals played an important role in preserving a trace of the dead after the war. Case studies frequently mentioned children who received the name of dead kin during the war, but without having the means to assess how widespread and how specific to WWI this behavior was.
Our finding is that in early twentieth-century France, even in peacetime, the death of a father during pregnancy led to a sharp increase in the probability that his offspring would carry his name. The traumatic nature of these events was apparently sufficient to elicit a sharp increase in such transmissions. During WWI, paternal name transmission, already boosted in the general birth population by the risks that soldiers relocated to the front incurred, was even more pronounced for prenatal orphans. In such cases, the practice assumed an additional meaning, since the namesake honored not just a father but also a patriotic hero. The wartime “Social Announcements” section of the conservative, upper-class-oriented newspaper Le Figaro frequently reported births that stressed that a father had “gloriously died for France.”
The study of uncle to nephew/niece transmission enables us to disentangle the effect of trauma from the cult of a fallen hero. The names of uncles who died during the war were transmitted to nephews and nieces for a consistently longer time than were those of uncles who died in peacetime. During the war and after the Armistice, the French population relied heavily on first names to keep the memory of fallen soldiers alive.
The upshot is that the practice of naming offspring for kin, a weakening anthropological tradition, received renewed life in early August 1914 when families were suddenly confronted with the horrendous prospect that they could never have anticipated. For families plunged headlong into mourning, naming a child for a father or uncle slain in war was a tribute that happened to comply with the patriotic attitude of the time. This study highlights that anthropological stability and behavioral fluidity are by no means contradictory. Faced with an unexpected change in death risk, parents collectively modified their behavior. What appeared to be a constant anthropological pattern, the naming of one child out of ten with the father’s name, turned out to be extremely sensitive to context: Parental preferences, customs, and habits can change from one day to another. A social theory of action should be able to take into account both the durability of habits and the fluidity of practice.
For the traditional naming for kin practice in Western Europe, see Agnès Fine, “L’héritage Du Nom de Baptême,” Annales ESC, IV (1987), 853–877; Ólöf Garđarsdóttir, “Naming Practices and the Importance of Kinship Networks in Early Nineteenth-Century Iceland,” History of the Family, IV (1999), 297–314; Frans Van Poppel et al., “Naming for Kin and the Development of Modern Family Structures: An Analysis of a Rural Region in the Netherlands in the Nineteenth and Early Twentieth Centuries,” ibid., 261–295. For baby names as reactions to historical events, see Michael Wolffsohn and Thomas Brechenmacher, “Nomen Est Omen: The Selection of First Names as an Indicator for Public Opinion in the Past,” International Journal of Public Opinion Research, XIII (2001), 116–139; Elena Obukhova, Ezra W. Zuckerman, and Jiayin Zhang, “When Politics Froze Fashion: The Effect of the Cultural Revolution on Naming in Beijing,” American Journal of Sociology, CXX (2014), 555–583; Martin Saavedra, “Kenji or Kenneth? Pearl Harbor and Japanese-American Assimilation,” Working Paper 8547 (Oberlin College, 2018), available at https://dx.doi.org/10.2139/ssrn.3218512; Robert Urbatsch, “The American Public’s Attention to Politics in Conflict and Crisis, 1880–1963,” Journal of Interdisciplinary History, XLVI (2015), 225–244.
For early works, see Jean-Jacques Becker, 1914: Comment Les Français Sont Entrés dans la Guerre Contribution à L’étude de L’opinion Publique, Printemps-été 1914 (Paris, 1977); idem, Les Français dans la Grande Guerre (Paris, 1980); for more recent works, idem, “Willingly to War: Public Response to the Outbreak of War,” 1914–1918-Online: International Encyclopedia of the First World War (2015), available at https://encyclopedia.1914-1918-online.net/article/willingly_to_war_public_response_to_the_outbreak_of_war; André Loez, 14–18, Les Refus de La Guerre: Une Histoire Des Mutins (Paris, 2010).
For early work on first names during the war, see Édouard Lévy, “Les Prénoms de La Guerre (1914–1918),” Revue Internationale de Sociologie, XXVI (1918), 508–513; idem, Manuel Des Prénoms (Paris, 1922); for the analysis of paternal name transmission to orphans, Todd, Alain-Jacques Valleron, and Pierre Bougnères, “The Naming of Orphans in France during World War One: A Study of a Nationwide Cohort of Pupilles de La Nation,” Historical Methods, LI (2018), 82–91; for classic references to the cult of fallen soldiers and mourning practices during the war, see George Lachmann Mosse, Fallen Soldiers: Reshaping the Memory of the World Wars (New York, 1990); Stéphane Audoin-Rouzeau and Annette Becker, 14–18, Retrouver La Guerre (Paris, 2003).
Geneanet is available at https://www.geneanet.org/.
R Core Team, R: A Language and Environment for Statistical Computing (Vienna, 2019); Murat Sariyar and Andreas Borg, “The RecordLinkage Package: Detecting Errors in Data,” The R Journal, II (2010), 61–67. All computer codes needed to replicate the de-duplication process can be found at https://github.com/nptodd/NamingForKinWW1.
Not until July 1915 did the French army organize a leave system that allowed soldiers occasionally to spend a few days at the rear. For a detailed analysis of the mechanism behind the May 1915 plunge, see Todd, “Long-Term Health Effects of World War I Stresses,” unpub. Ph.D. diss. (Pierre and Marie Curie Univ., 2017), 69–76.
For a recent use of crowd-sourced genealogies in population genetics, see Joanna Kaplanis et al., “Quantitative Analysis of Population-Scale Family Trees with Millions of Relatives,” Science, CCCLX (2018), 171–175; for the use in social history, Arthur Charpentier and Ewen Gallic, “Using Collaborative Genealogy Data to Study Migration: A Research Note,” History of the Family, XXV (2020), 1–21; idem, “La Démographie Historique Peut-Elle Tirer Profit Des Données Collaboratives Des Sites de Généalogie?” Population, LXXV (2020), 391–421; Tine De Moor et al., “Long-Term Trends in Marriage Timing and the Impact of Migration, the Netherlands (1650–1899),” Historical Life Course Studies, VI (2017), 40–68; for resistance to state policy in relation to names, James C. Scott, John Tehranian, and Jeremy Mathias, “The Production of Legal Identities Proper to States: The Case of the Permanent Family Surname,” Comparative Studies in Society and History, XLIV (2002), 4–44.
Becker, “Willingly to War: Public Response to the Outbreak of War.”