1“It is a dark, deep, and continuous stream of mortality which has been running through our statistics since the commencement of registration – indeed, from the most remote periods – without showing any signs of abatement – and occasionally overflowing its banks, producing epidemics with most disastrous results.
2It is a mortality which is to a great extent preventable, as evidenced in the enormous improvement which has of late years marked the records of lying-in institutions generally and under conditions where precautionary measures are strictly observed. Then if preventable, why not prevented?” (Williams, 1904, 1.)
3With these words, Dr William Williams, a Welsh medical officer, lecturer at Cardiff University, opened his Milroy lecture on “Deaths in childbed. A preventable mortality,” delivered at the Royal College of Physicians, in London in 1904 . Dr Williams shows a considerable awareness of the long history of maternal mortality and the halting progress, if any, that had been made. Yet progress was possible, if only the right preventative measures were taken. His is a cry for effective intervention. No doubt his Welsh background had conditioned him to see the problem: maternal mortality had persisted at remarkably high levels in Wales. As Irvine Loudon acutely observed: “Maternal mortality was the great exception in what has been called the great mortality decline. From the end of the nineteenth century until the mid-1930s, when death rates in general and infant mortality in particular were steadily declining, maternal mortality scarcely altered (Loudon, 1992b, 5).” Indeed, he believed that in some places it had actually increased, though as we discuss later there are some issues about the reliability of the medical data.
4The later nineteenth and the early twentieth century witnessed the rise of a maternal and infant welfare movement in response to three social and demographic problems. These were high maternal mortality, high infant mortality and falling marital fertility in many parts of the western world. War and empire were catalysts for change. The Boer War and on a much greater scale, the First World War, decimated populations, particularly affecting the younger age cohorts of males. There was need to reduce the wastage of lives and encourage population growth for national and imperial reasons. But there were also humanitarian motivations at play that centred on the desire to save the lives of mothers and young children (Fildes, Marks, Marland, 1992, 1).
5There was also a shift of consciousness. Thanks to the development of social medicine and more active interventions by states in the provision of health care, as well as the work of international health organisations, many health problems came to be perceived as unacceptable and intolerable. This was particularly true of high infant mortality, which was a scourge common to all countries, albeit to varying degrees (Pozzi, Barona, 2012; Galiana-Sánchez et al. in this issue).
6Still, it was only after the First World War that a widespread awareness of the need to tackle maternal and infant mortality emerged in Western countries. Politicians, religious leaders and health authorities, perhaps for different motives, became deeply concerned about the decline in the birth rate. This served to focus attention on maternal health and the implications for the care and survival of infants. Woodbury observed that the growing interest in maternal mortality was also driven by “a realisation that a large proportion of the mortality and sickness caused by pregnancy and confinement is preventable” (Woodbury, 1926a, 1). However, a reading of the medical articles of the time makes it clear that the initial impetus came from concerns for the health of the babies rather than the mother.
7“Preventable” and “preventability” are words commonly repeated in texts on childbirth mortality in the past, and even today. This determination to save lives found expression in a seminal initiative undertaken by the Public Health Relations Committee of the New York Academy of Medicine which carried out a study of maternal mortality in New York City, coordinated by Ransom S. Hooker (1933) . A Committee, including four obstetricians, analysed all the maternal deaths which had taken place in New York City in 1930-32 and not only established how many maternal deaths were preventable, but also identified who was responsible for each death. Was it the attendant, the medical doctor or midwife, or was it the patient herself?
8The Committee classified two out of every three deaths as preventable (1,343 out of 2,041 cases). In its view, if there had been proper treatment and care, these women could have been brought safely through parturition (Hooker, 1933, 32). In assigning responsibility for the preventable deaths, the Committee found that “sixty per cent of all the deaths which could have been avoided have been brought about by some incapacity in the attendant; lack of judgement, lack of skill, or careless in attention to the demands of the case… more than a third of all preventable deaths were due to some failure on the part of the patient herself to take advantage of those facilities which are at hand for safeguarding her in the period of gestation and lying-in. This element in the situation is one of education entirely (Hooker, 1933, 49).” The Committee added the wry comment that pregnancy was of such frequent occurrence that many mothers did not look on it as a condition that merited any particular consideration.
9This was the most detailed and in-depth study to date, but other reports published around this time reached broadly similar conclusions with regard to the proportion of maternal deaths that were preventable. For instance, the Report of the Maternity Service Committee on Maternal Mortality and Morbidity in Northern Ireland published by the Ministry of Home Affairs of the Government of Northern Ireland in 1943 noted that “intensive investigations were previously carried out in England and Wales by a departmental committee, in Scotland by officers of the Board of Health, and in America by officers of New York Academy of Medicine. The conclusions of these committees were the same, namely, that at least half of the maternal deaths were preventable  (Government of Northern Ireland, Ministry of Home Affairs, 1943, 7).” However, a report on maternal mortality in Wales a few years earlier estimated much more conservatively that one-in-five maternal deaths might have been prevented had there been reasonable care and attention during pregnancy (Ministry of Health, 1937, 99). Indeed, the authors of the report, in a section headed “The mother’s responsibility,” were of the view that “a large proportion of the deaths considered preventable” were due to “neglect by the mother herself” (Ministry of Health, 1937, 100).
10While attribution of responsibility might vary, the emerging pattern from these various studies adds force to Williams’ rhetorical question: “Then if preventable, why not prevented?” One might wonder why it took so long to eliminate these avoidable tragedies and indeed why they persist to this day in some poorer countries. According to recent data from the World Health Organization, every day, approximately 810 women die – most of the deaths (94 %) occur in low and lower middle-income countries – from preventable causes related to pregnancy and childbirth .
11An element in the “apparently intractable issue of high rates of maternal mortality” (McIntosh, 1997, 1), both in the past and in some communities nowadays, seems to be an acceptance of the inevitability of childbirth mortality, which is sometimes linked to a sense that there is a price which must be paid for motherhood . A similar sense of resignation had long surrounded high infant mortality  but these attitudes were under pressure by the beginning of the 20th century. We also have to take into account the dominant discourse around motherhood. Before the First World War motherhood “was elevated to the status of an almost religious vocation, with ideas of ‘duty’ and ‘sacrifice’ abounding in any discussion of mothers.” Speaking at the London Conference on Infant Mortality in 1906, John Burns, president of the Local Government Board (forerunner of the Ministry of Health) famously commented that people should “glorify, dignify and purify motherhood; for what the mother is, the children will be” (McIntosh, 1997, 97). Another commentator of the US Children’s Bureau spoke of death while giving birth as a serious loss to the country. “The women who die from this cause are lost at the time of their greatest usefulness to the State and to their families; and they give their lives in carrying out a function which must be regarded as the most important in the world (Meigs, 1917, 9).” We return to the idea of maternal sacrifice in more depth later in the context of the harrowing choice sometimes faced between safeguarding the mother’s life or the life of the baby.
12Attempts at improvement, as we have noted, were initially focused on saving infant lives. Gradually the health of the mother came into the picture but mainly indirectly and instrumentally . There was a dawning realisation by medical practitioners and others of the “far-reaching influence over infant mortality that is exerted by the health and condition of the mother,” as exemplified by the sentiments of the medical doctor, Robert Morse Woodbury  (1926a, 1). This implied greater attention should be paid to the provision of antenatal and postnatal services. Woodbury, it is worth noting, had been the director of statistical research in the Children’s Bureau in the Labor Office in the United States, which was particularly active in the early decades of the 20th century in the field of infant and maternal welfare. The Bureau had published several reports on maternal mortality . Thus, there was only a limited focus on the health of the mother herself during pregnancy and childbirth. Antenatal care, where it did sporadically exist, was for the welfare of the unborn child, not its mother, and postnatal care did not exist (McIntosh, 2012). This relates to Britain but it could be generalised to many Western societies.
13Authorities in the United States paid special attention to infant and childbirth mortality in the opening decades of the 20th century, a concern driven in part by unfavourable comparisons with other countries . A raft of reports highlighted this dismal record . The American reports offered rich and detailed information not only on the United States but also comparative data on European countries and acute observations on the quality and comparability of the international statistics .
14To dramatise the problem of infant mortality the Children’s Bureau of the US Department of Labor (1921) made use of a kind of pictogram, featuring a thermometer that showed different levels of maternal mortality rates for different countries . This showed that during the First World War the United States was the worst performer, followed by New Zealand, Scotland and Australia, out of a selection of 17 countries. By the end of the Second World War the relative situation in the USA had improved considerably, though as table 1 shows there were still major contrasts between different countries in the 1940s.
Maternal mortality thermometer in 17 countries in 1919
Maternal mortality thermometer in 17 countries in 1919
15Looking across the countries and regions for which data exist, the geographies of maternal mortality on the one hand and infant mortality (or more specifically neo-natal mortality) on the other do not readily match up. This is surprising as one might expect a closer link between maternal mortality and neonatal mortality because the latter is bound up with events during pregnancy and childbirth. By contrast, the causes of post-neonatal mortality tend to be exogenous, and caused mainly by respiratory or gastro-intestinal diseases.
16Before leaving the issue of data gathering and quantification it is helpful to note its special significance within the history of maternal health reform. Those countries that enjoyed an early reduction in maternal mortality were also those that collected numerical information on the problem. This was so in the case of Sweden (1749), Denmark (1801) and Norway (1801). In the Netherlands vital statistics only became available in 1839 but the reduction of maternal mortality was similar to that in Sweden. In other countries such as Belgium (1830), England and Wales (1838), and Germany (1841), information on levels of maternal mortality only became available late in time. Admittedly, information alone was not enough and the willingness of governments to improve the general health status of the population was a prerequisite for reform. The availability of relevant data, the larger picture as it were, also proved vital to the activities of pressure groups in Britain and America who were fighting for reform in the early 20th century (De Brouwere, 2007, 555).
Infant mortality thermometer in 1919
Infant mortality thermometer in 1919
Infant and maternal mortality rates (per 1,000 live births) in selected countries in 1919-21, 1929-31, 1938-40 
Infant and maternal mortality rates (per 1,000 live births) in selected countries in 1919-21, 1929-31, 1938-40 
17While mother and child health welfare programmes were in operation in many western societies from the period of the First World War onward, it is still striking that for long there was relatively little attention devoted to the health of the mother. As if history was repeating itself, this time in the developing world, an article provocatively titled “Where is the M in the MCH?” – the reference is to the “mother” in Mother and Child Health programmes – was published in 1985. The neglect of maternal mortality in the poorest countries was still evident then and later. In 2015 more than 300,000 women died of complications during pregnancy or up to six weeks after giving birth, according to the World Health Organization, and this was down from a half million maternal deaths in 1990 . These tragedies were to be found overwhelmingly in the poorer countries of the world.
18This is just one example that shows that problems, replete with tragic consequences, can recur in the fields of maternal and infant mortality if earlier historical experiences are ignored or forgotten. The mainspring for this special issue of the journal comes from a research network that seeks to draw lessons from the past and apply them to contemporary contexts. More specifically, it seeks to combat poverty-related diseases affecting maternal and infant health. A good understanding of earlier debates and practices, we would argue, helps inform more effective strategies in the present. As part of the activities of this network, a session at the third conference of the European Society of Historical Demography was organised in Pécs, in June 2019. The session was mainly devoted to maternal mortality in Europe in the 19th and 20th centuries and included also the link to perinatal health . Four of the six articles in this special issue of the journal are refined and extended versions of papers presented at that session.
19Referring back to Williams’ lecture and his clarion call for action, clearly he felt it was high time that serious attempts were made by the legislature, the medical profession, and others “to put an end to such a painful sacrifice of mothers which stealthily but surely takes place annually” (our italics) (Williams, 1904, 3). There is a suggestion here in the use of the word “stealthily” that some, perhaps many maternal deaths were going unrecorded. Fast forward to more recent times and we find a more explicit statement of this likelihood. “If dead women are not even counted, then it seems they do not count. We have an invisible epidemic,” said Joy Phumaphi, the World Health’s Organization Assistant Director-General on Family and Community Health, at a meeting which took place in 2004 in Nairobi .
Issues, methods, debates
20The call for papers for the Pécs conference was virtually a manifesto of what is needed by way of research. It is also clear that further investigations are warranted, including the use of individual-level data to supplement aggregate-type studies based on official statistics. The call for papers was structured under some but not all of the headings shown below. What these indicate is the scope for an expanding agenda for research.
21It has been shown that relying uncritically on official statistics can be misleading. In comparative studies there is the problem that the criteria for the recording of maternal deaths and the classification of causes of death may vary between countries. The efficiency of the different authorities in collecting and tabulating data may also vary. But these potential limitations also apply within the same country, even when making subregional comparisons over time, as is demonstrated by the cases of Ireland and Northern Ireland. Pozzi and Kennedy in this issue show that comparisons between different regions of the island are of doubtful validity, though Dublin-Belfast comparisons seem to be more soundly based. Dublin appears to have performed better in terms of limiting maternal mortality than Belfast, though the rates in both cases were high by comparison with England, for instance. The greater availability of trained midwives and deeper traditions of institutional maternity care in Dublin help account for the difference.
22An alternative strategy, that of using individual-level data, allows us to explore aspects that we just cannot analyse by relying solely on official statistics. This is evident in a number of contributions by Alice Reid (Reid, 2012; Reid and Garrett, 2018) where individuals are the unit of analysis and the findings illuminate a range of possibilities beyond the reach of aggregate-type studies. In particular, instances of hidden maternal mortality are revealed.
23However, before throwing the baby out with the bathwater it is as well to recognise that studies based on individual cases also have their limitations. Unless conducted on a massive scale, which is unrealistic in terms of most research budgets, these micro-studies do not allow us to establish patterns and trends for large populations and territories. Moreover, the measures of maternal mortality are sometimes of an indirect kind. In Ireland, Britain and some other countries the cause of death is indicated clearly enough but in others, Italy being a prime example, a specific cause of death may be lacking. Moreover, maternal deaths following a stillbirth or miscarriage may be difficult to identify which is a challenge to nominative record linkage.
24In view of the strengths and limitations of the two methods, there is an obvious invitation to compare the results from the analysis of aggregate data with those found from nominative reconstructions of small populations for the same areas. This has been done by Manfredini and Breschi in this issue and they find a reassuringly high degree of consistency between the maternal mortality rates derived from the Italian official statistics for some Italian regions and those based on the nominative reconstitution of small populations for the same regions.
The association between socio-economic status and maternal mortality
25Perhaps surprisingly, the literature on the connection between socio-economic status and maternal mortality is not consistent across different European countries, regions and time periods. Why this might be the case is not at all clear, so there is a major research agenda to be pursued here. Loudon has suggested the idea of a “reverse social class connection.” This is on the basis of local studies of English maternal mortality, as well as official publications on MM ratios of married women classified according to social class by the Registrar General of England & Wales (1936, 131) in 1930-32. As might be anticipated, most infectious diseases and also infant mortality during the 19th and 20th centuries affected the working classes most of all, and not just in Britain. Mortality rates were highest in the working classes and lowest in the upper classes. But as regards maternal mortality, according to Loudon, the reverse turns out to be the case.
26Common sense suggests that women of the lower social classes were more likely to be badly nourished and anaemic, and therefore to have less resistance to the various causes of maternal mortality. However, social class was a much weaker determinant of mortality than the type of birth attendant. The higher mortality of women in social classes I and II (England and Wales 1930-32) was due to their being delivered by doctors, mostly general practitioners, who were much more likely to undertake repeated vaginal examinations and use instruments in normal labour. Midwives who mostly attended mothers in social classes IV and V interfered much less. There is another important point: general practitioners were much more likely than midwives to come into contact with non-maternity patients suffering from streptococcal disease and so became carriers of the bacterium. Unless they were scrupulous in antiseptic and aseptic procedures in midwifery (and many were not) they were much more likely than midwives to infect their midwifery patients (Loudon, 1992b).
27This line of interpretation is highly plausible, and yet some doubts must remain as to the extent to which this might be generalised in time and space. It would be helpful to have more in-depth analyses based on individual-level data so as to arrive at a fuller understanding of the mechanisms linking socio-economic status and maternal death in childbirth. Alice Reid’s contribution (this issue) illustrates not only the role of midwifery in the decline of maternal mortality but also an urban-rural dichotomy. In rural areas we have to bear in mind the distances that might have to be travelled and the implications this might have for maternal and infant mortality. These are further promising areas of enquiry.
28A link between poverty, nutrition and maternal mortality has been attested to by many medical practitioners down the years, as indicated by Pozzi and Kennedy (this issue) for Britain and Ireland. But, on reflection, the relationships may not be of the simple linear kind. Recent advances in medical, nutritional and biochemical studies have given rise to renewed interest in this area and the eventual findings may prove to be more complicated than originally thought.
The link between maternal and perinatal / neonatal mortality
29Once again and against expectations, the relationship between maternal and neonatal mortality has been brought into question, at least for some countries during the first half of the 20th century. The systematic study of individual-level cases and associated socio-economic variables should prove illuminating here. The interrelationship between maternal mortality and fertility also needs to be brought into the mix.
30The American statistician Robert Morse Woodbury (1926b, 43) found a strong relationship between maternal death and the chances of survival of the infant. “Infants whose mothers died within one year following the confinement appeared to be subject to a considerable handicap, for their mortality rate, 450.0 per 1000, was over 4 times as high as that for other infants, 109.2. In those cases in which the mother died within one month of confinement the mortality was even higher. Six in every ten of these babies died before the end of the first year.” Similar results were reported by Yerushalmy et al. (1940).
31A close connection between the two mortalities might seem to be a matter of common sense. But Loudon thinks otherwise: “I suggest that a close link implies that infant and maternal mortality rates responded in very much the same way to a number of well-recognised determinants of mortality. Examples of such determinants would be income-levels, social class, the quality of the environment […] the effects of living in urban or rural areas, nutrition, parity, sibship size, maternal age, and, of course, the quality of maternal care. To talk of ‘close links’ also implies that we should expect to find similar secular trends, and similar levels when the rates in different countries and regions are compared (Loudon, 1991, 30).”
32Using time series data on MMR and IMR for a large sample of countries and regions, Loudon came to the conclusion that there was no necessary link between MMR and IMR. The explanation seemed to be that most maternal deaths were due to causes which did not put the neonate’s life at risk. What were these? In his view, maternal mortality was primarily determined by the quality of care provided by birth attendants. He decried poor obstetric care resulting either from ignorance of basic procedures or unnecessary interventions by those attending births (Loudon, 1991, 72). Thus, maternal mortality was relatively insensitive to social and economic factors except in so far as these determined the type and quality of the birth attendants. Mothers were at risk from the two ends of the spectrum: cheap and untrained midwives could be a menace but so also could expensive and overzealous doctors.
33It is possible that Loudon may have overplayed his hand. Levels of neonatal mortality, it would seem, were determined to a large extent by factors we do not fully understand. The quality of maternal care certainly played a part but may be exaggerated. But Loudon is surely correct in making a distinction between the consequences of the death of the mother for the health of her child – interruption to breastfeeding and the absence of close maternal care – and the quite different issue of possible links between maternal mortality and infant mortality. The two situations may well not share common determinants. Even so, we might hypothesise that intensive studies of individual cases, depending on time period and context, might show up statistical and more material links between maternal mortality and infant mortality. Worth exploring also, we suspect, are the possible interrelationships between prematurity and congenital deformities, maternal death and environmental conditions.
Medical science, technology, midwifery and the hospitalisation of childbirth
34Emmanuelle Berthiaud (this issue) provides the essential historiographical context in which science, technology, religious and professional practices all interrelate. The temporal span is wide, as she documents developments in infant and maternal welfare from the 16th century through to the 1870s. France is the centrepiece but she also takes account of medical knowledge in the United Kingdom, Germany, Spain and Italy. This geographical spread is in itself significant as she finds evidence of the widespread circulation and exchange of scientific ideas internationally. The 18th century, it appears, marked a real break in the history of birth and motherhood. The medicalisation of childbirth was in its infancy, in some places obstetricians with new instruments and procedures attended the birth scene, hospitals for pregnant women were coming into being, and the first obstetrics schools were established. These developments coincided with profound changes in conceptions of childbirth, images of the female body, and its reproductive function (Gélis, 1988).
35For particular countries and time periods it has been documented that ordinary midwives assisting with home deliveries ensured lower perinatal and maternal mortality than medical doctors in lying-in hospitals, while in other countries and time periods the results are quite different. One reason presumably is that the role of midwives varied greatly between countries (Marland and Rafferty, 1997). Marland and Rafferty provide a panoramic overview of this diversity in practices. In Austria, Hungary and Switzerland midwives were severely limited in what they could do. If there were any complications with the delivery, they were obliged to call in a doctor. In Germany, however, their role encompassed a wider range of responsibilities. In emergencies where they detected signs of eclampsia or placenta previa in the pregnancy they could decide to take the mother to a clinic without first consulting a doctor. German midwives were permitted to carry out vaginal and rectal examinations, having first disinfected themselves and put on rubber gloves. These examinations had to be recorded and the case notes presented to the town doctor. Under certain circumstances German midwives were allowed to deliver twins and handle breech births. They were not permitted, however, to administer medicine that would stimulate contractions or relieve pain. In England the midwife was able to give pain relief in some instances. In Belgium the midwife could give injections of ergotine, caffeine and camphor, and administer opium tincture. The use of instruments, however, was not seen as being within her capability. In more liberal Sweden midwives were empowered to use obstetrical instruments. Swedish midwives were also trained to repair ruptures, as were their French counterparts. French midwives also performed vaccinations against smallpox and tuberculosis (Marland and Rafferty, 1997, 1).
36Högberg (2004) and De Browere (2007) have shown the relevance of effective collaboration between medical doctors and midwives  and of a mutual appreciation of the skills and competence of the two professions. In similar vein, Loudon (2002) has argued that the success of midwives was dependent not only on effective training but also on being accepted and respected as professionals by the communities they served, and preferably by the medical profession itself . The last is a somewhat elusive variable: Loudon (2002) and also Breathnach (2015) touch on the different degrees of acceptance enjoyed or suffered by midwives in different countries.
37By the 1920s there was a widespread recognition by British and American obstetricians that the performance of north European countries such as the Netherlands, Denmark and Norway was well in advance in terms of limiting maternal mortality . Some put this down to low levels of industrialisation and the cleanliness of these countries. Others even claimed that the differences were racial in origin. Most would agree that there are issues about the quality of the midwifery services, the nutritional status of the mother, and a host of other variables that might account for the differing findings. Controlling for the relevant variables is the challenge. In this context one might note the findings of Alice Reid for England and Wales (this issue). Taking industrial structure as well as background mortality into account, she found that the training given to midwives had a favourable impact on maternal mortality.
38The case of Valencia, analysed by Eugenia Galiana-Sánchez et al. (this issue), illustrates some of the general points made earlier. The Spanish city of Valencia was characterised by high rates of maternal and infant mortality, as was the case elsewhere in the country. But a growing awareness of the problems in the early decades of the 20th century led to vigorous attempts at reform. Particularly significant was the establishment of the Escuela de Puericultura of Valencia (“School of Child Care”) and other health initiatives, such as the Asilo de Lactancia (“Infant Nursery”) and the Gota de Leche (“Infant Milk Depot”). It is noteworthy that these exemplary initiatives began with a focus on infant rather than maternal welfare. But we may also note that maternal education, whose motivation was to ensure healthy pregnancies and reduce congenital problems, promoted maternal health at least indirectly.
39The hospitalisation of childbirth could also present some drawbacks, depending on time period. Manfredini and Breschi show (this issue) that part of the regional differences within Italy in performance towards the end of the nineteenth century can be accounted for by the larger share of home births in southern Italy. Paradoxically, home births preserved mothers from the high risk of puerperal infections that were so common in the public and private hospitals of that time.
40There is a fascinating difference between, roughly speaking, northern and southern Europe in terms of male midwives attending childbirth. According to Filippini (2020), male midwives or “accouchers” began to appear in the late seventeenth and early eighteenth centuries in northern France, Great Britain and the Netherlands. This trend continued into the nineteenth century and extended to other countries, including Germany. But almost nothing of this kind was recorded for southern Europe, where childbirth assistance remained firmly in the hands of midwives throughout the 19th and the first half of the 20th century. Gender mattered, it is clear, but in different ways in different parts of Europe. Filippini (2020) argues that these geographical differences were due to differences in natural conditions, culture and religion. Some exceptions apart, in southern Europe it was considered indecent for a woman to be visited by a man in the intimate circumstances of giving birth. Moreover, expectant mothers seem to have had a lot of confidence in the person of the midwife. She in turn was resistant to calling on a doctor for assistance when a birth was difficult, as this was considered an admission of defeat. And then there was religion.
41The geographical limits to the spread of the accouchers correspond fairly closely to the boundaries of the Protestant regions of Europe. In these regions a different ethic prevailed, one that dispensed with traditional notions of “decency” in favour of considerations of health and survival. In Catholic countries such reforms were inhibited by the persistence of fatalistic attitudes towards death and also by the persistence of values and symbols of maternal sacrifice. The role model for female virtue and chastity, and for motherhood, was the Virgin Mary, the Mother of God. This feminisation of the Christian message laid the basis for Church control of female sexuality, which in turn meant that intrusions by state and science into family life were resisted. As night follows day, conflicts between church and state, between science and religion, proved inevitable. But a more nuanced reading might explore how the principal agents – bishops, priests, nuns, politicians, medical doctors and other professionals – navigated these troubled waters and effected makeshift compromises. Italy and Ireland are particularly rich in possibilities .
42Within the vast process of radicalisation of childbirth and the development of modern obstetrics in the western countries, the introduction of medical practices such as therapeutic abortion, caesarean section, and embryotomy command particular attention. These operations were the surgical options that obstetricians had at their disposal in cases of obstructions during childbirth in the 18th and 19th centuries. Sometimes obstetricians were faced with a terrible dilemma, the choice between the life of the mother and the life of the unborn child.
43Caesarean section was the only operation which could, at least in principle, save both lives. In practice, the mother usually died from peritonitis infection and internal haemorrhaging. Induced abortion, as well as embryotomy, offered the woman a better chance of survival, but caused the immediate death of the unborn child. Other procedures like labour induction before full term could result in live birth but these premature infants rarely survived for long outside the mother’s womb. Advances in medical technology, such as these, gave rise to heated ethical debates, a dialectic that continues to this day. Religion was the great source of ethical values and judgements, so we find marked differences between countries with different religious affiliations. In Catholic countries like France, Belgium and Italy, for instance, caesarian sections were common whereas in countries like Britain and the United States this operation was unusual and other interventions were preferred (Filippini, 2020).
44Belgium is a particularly interesting case in this respect. Though a Catholic country, it had a liberal constitution since its foundation in 1831 and the medical profession was roughly equally divided between Catholic and liberal doctors. Gijbels (this issue) traces these ethical debates across the decades between 1840 and 1880 and shows how religion dominated, even though some doctors wished to limit the discussion to more purely medical arguments. Medical data indicated that in the choice between embryotomy and caesarean delivery, the latter was associated with high maternal mortality. Despite this, Catholic doctors opted for the riskier caesarean delivery on moral grounds. Thus, the choice of surgery was dictated by religious belief rather than considerations of risk to the mother’s life.
45These “moral dilemmas” drag a range of other questions in their wake. Among theologians, some spoke of the natural propensity of women to maternal sacrifice. Even Eve’s responsibility for original sin was dragged into the reckoning. Other theologians believed the life of the mother took precedence over that of the foetus. Others still made subtle distinctions between killing – an active intervention – and allowing death to ensue as a byproduct of a therapeutic intervention whose intention was not one of killing.
46Not only religious but demographic and eugenicist arguments intruded into the debate and might operate in combination. If the choice lay between saving the life of an undernourished mother, perhaps afflicted by rickets or some other deformity, and that of a potentially healthy child, the decision was clear-cut. The child was more valuable to society. The woman in fact was failing in her family and social duty to bear children. As the Italian obstetrician Luigi Pastorello wrote: “I do not know therefore whether society cares more about a healthy and well-shaped child, such as usually can be saved by caesarean section, or about a woman unsuited to such purpose” (Pastorello, 1838, 81-82 as quoted in Filippini, 2020, 348). Some doctors even made calculations as to the lower demographic loss if the mother’s life was sacrificed.
47The moral dilemmas didn’t end there. There is the question of who had the authority to decide between the life of the mother and that of the child. Was it the doctor, the husband or the wife herself? These are the kinds of questions that still beset current bioethical issues. Some commentators saw doctors as natural judges and super partes. What is really striking is that few felt the mother herself should have agency. Some felt she should not even be informed of the mortal risk she faced. The mother was simply the object of discourse (Filippini, 2020).
48Wealth and social class might well determine the actual outcome. A doctor paid for by the family would surely consult the family or at least the husband and father of the child. But when it came to a poor woman in labour there might be no such consideration. Alfonso Corradi (1874-1877) collected information on 158 cases of caesarean sections performed between 1780 and 1876 in Italy. Corradi calculated a 66 % maternal death rate (105 out of 158) from these operations . Women from wealthier backgrounds were in a position to avoid such fates through the use of therapeutic abortion or induced premature labours (Filippini, 2020).
49In Catholic Europe religion influenced, often dictated the kind of surgery performed. Viewed over the longue durée, Church understandings of the foetus, of ensoulment, of pregnancy and birth did change. In ancient and medieval times, under the influence of Roman law, the foetus was considered pars viscerum matris (part of the mother’s body) and did not possess an individual identity of its own. So, a therapeutic intervention aimed at saving the foetus would have made little sense. Many of the methods used for contraceptive purposes in fact resulted in induced abortions .
50For centuries theologians had supported the mother’s right to defend her life, viewing the foetus as an involuntary aggressor (Filippini, 1995, 2020). But from the 16th century onward, theologians could be found who argued for the priority of the life of the child. It became the duty of the mother to sacrifice herself in favour of the child under pain of losing her own hope of eternal salvation. However, the Magisterium (the Church’s teaching authority) moved slowly on the question. In 1852, when requested by a canon of the University of Leuven for a ruling, the Church decided not to answer (nihil esse respondendum, there is no answer). It entrusted the matter to local judgement, taking into account the specific circumstances of the case.
51It was only in the later 19th century that the Church officially intervened with explicit pronouncements from the Holy Office (Sant’Uffizio). A series of pronouncements in 1884, 1889, 1895 and 1902 set limits to obstetric interventions. The Church prohibited any operations that directly or indirectly adversely affected the life of the foetus. The stance was now one of defending the life of the foetus. With a subtle semantic shift this became a “defence of life,” and any form of abortion was condemned. The life of the mother became an oblation or a sacrifice offered up to God (Filippini, 2020, 352-353).
The status of women within the family and the wider society
52This capacious theme embraces so much. It is surely necessary to bear in mind the social and economic role of women, and its evolution over time. Earlier on we commented on the marginal attention bestowed on pregnant women, by contrast with the concern directed at infant welfare. This brings into the reckoning not only medical practitioners, midwives, hospital administrators and policy makers but also the couples’ relationship and even wider kin relationships, including mothers and mothers-in-law. The influence of societal forces acting on the family, from legislation relating to women’s work and education through to cultural expectations about women’s place in society, must also find its place within any comprehensive explanatory sketch.
53Education for parenthood was a common plea. But this could sometimes veer dangerously in the direction of eugenics. The woman, it was announced in one American publication, was “the temple of the life of this world to come.” The author continued in equally grandiose terms: “She must honor and care for herself accordingly; and this twofold aspect of her present and future duty, in caring for herself and in choosing her co-creator of the future, must be instilled into her mind with the solemnity, the sanctity, and the authoritative sanction of a religious dogma (Saleeby, 1911, 36).”
54This quotation comes from an article in the journal of the American Statistical Association of all places. It is titled “The mother and infant mortality,” where the author after giving some standard advice for mothers, wrote: “Education for parenthood should include more than instruction in the science and art of infant and child care, or in home management or even in the ideals of parenthood; it should also include instruction, direct or indirect in ‘the selection of parents’ (Hibbs, 1916, 78).”
55All this flowed easily from the assumption, found among leading eugenicists that “in all times and places woman’s primal and supreme function is or should be that of choosing the father of the future” (Saleeby, 1911, 36). In effect, she should ensure that breeding was confined to the fittest of the race. How the discourse of eugenics intersected with discourses more specific to maternal and infant mortality, and associated anxieties about the quality and quantity of children, might be a fruitful and challenging area of further enquiry.
56Fiction has a habit of mirroring real life, conveying feelings and emotions that are not readily represented in more academic styles of writing. An awareness of the trauma of maternal mortality, and its inherent drama, is evident in the many works of fiction that depict deathbed scenes involving the death of a heroine while giving birth. Such scenes were common in Victorian novels, reflecting the social reality of the period, and they also find a place in American writings. The protagonist in Hemingway’s novel, Farewell to arms (1929), it may be recalled, is faced with the death of the woman he loves. She is haemorrhaging shortly after a caesarean section. Even the baby which shortly before he had almost accused of killing the mother was born dead.
57“Everything was gone inside of me. I did not think. I could not think. I knew she was going to die and I prayed that she would not. Don’t let her die. God please don’t let her die. I’ll do anything for you if you won’t let her die. Oh, God, please, don’t let her die. Please, please, please don’t let her die. God please make her not die. I’ll do anything you say if you don’t let her die. You took the baby but don’t let her die. That was all right but don’t let her die. Please, please, dear God, don’t let her die (Hemingway, 1929, 353).”
58But such tragic scenes, as we have seen for different countries, were only slowly and belatedly being transmuted into effective health-care provision for mothers-to-be. To generalise greatly, standing in the way of amelioration there was an extensive inventory of hurdles: misogyny, fatalism, religious world views, cultural practices that discriminated along gender lines, sentimental celebrations of motherhood invoking images of suffering and sacrifice, disputes among clergy, doctors and midwives, and the conservatism and myopia of medical practitioners and policy makers. These variables, or subsets thereof, combined in different ways in different countries to produce uneven timelines for reform. There is a large research agenda here. Nor is it one that belongs exclusively to the past. Maternal mortality is still a major killer in many parts of the world. The social role of the historian kicks in here. Distilling lessons from history that might better inform policies in contemporary developing countries needs to be based on robust research findings. This volume is intended as a contribution to that socially desirable objective.
The quotation at the beginning of this section comes from the Milroy lecture given by Williams, in 1904. At that time the University still retained its original name “University College of South Wales and Monmouthshire.” The lectures have been given annually on public health topics since 1888. Their name derives from a Scottish physician Gavin Milroy, elected fellow of the Royal College of Physicians, who died in 1886 and whose legacy to the College financed a lectureship on subjects connected to “state medicine and public health.”
See Yerushalmy (1945, 140) who lauded this study as a “monumental contribution in this field.”
Pozzi-Kennedy in this issue.
The data refer to the year 2017. See www.who.int/news-room/fact-sheets/detail/maternal-mortality.
A Scottish obstetrician, John Martin Munro in 1933, wrote: “Maternal mortality and morbidity can never be reduced to zero… A price must be paid for motherhood.” See McIntosh (1997, 2).
These attitudes of resignation were common in Spain, where a ritual phrase “the little angel rejoices in heaven” was recited at the funeral of newborns (Bernabeu-Mestre et al., 2007).
See for example the article by Galiana-Sánchez et al. (this issue) on experiences in the city of Valencia in the early 20th century.
Woodbury, author of a report on maternal mortality published in 1926, spelled out this emphasis in some detail: “Interest in maternal mortality has been especially stimulated in recent years by the progress of the movement for reduction of infant mortality. A very considerable proportion of all deaths of infants under 1 year of age occur during the first month of life from causes which have their origin in the care and condition of mothers during pregnancy and confinement […] Reduction in the mortality from these causes depends upon improvement and extension of facilities for prenatal, confinement and postnatal care (Woodbury, 1926a, 1).”
A similar and even more explicit recognition of the fact that the welfare of the mother was subordinate to that of the child was contained in the Report of the US Children’s Bureau (Meigs, 1917). In the letter of transmission of the Report, Julia Lathrop, the Chief of Bureau on September 25, 1916 wrote: “This report has been prepared because the bureau’s studies of infant mortality in towns and rural districts reveal a connection between maternal and infant welfare so close that it becomes plain that infancy can not be protected without the protection of maternity.”
See also the “White House conference on children and youth,” a series of meetings (and associated publications) organised since 1909 for more than 70 years by the President of the United States, on issues relevant to the decade in which the Conference took place. See in particular the White House conference on child health and protection, November 19-21 1930, Washington.
See the publications by the Children’s Bureau of the US Department of Labour. We refer to some of these reports only.
See in particular Tandy (1935).
Children’s Bureau, US Department of Labor (1921).
Time periods are not identical for all countries and territorial boundaries in some cases changed over time. For more detail see Yerushalmy (1945, 135).
BBC news report, 12 November 2015: www.bbc.co.uk/news/world-europe-34796833.
The title of the session “Maternal and perinatal health in Europe in the 19th and 20th centuries” made the link between the two even more explicit. The research network is also linked to the Spanish project “Past and present in the control of neglected diseases of poverty: the historical example of Mediterranean Europe and international health cooperation” (HAR-2017-82366-C2-2-P), Ministry of Science, Innovation and Universities, Government of Spain.
www.who.int/mediacentre/news/releases/2004/pr65/en/. See also Horsbrugh-Porter (2009, 93).
Högberg (2004, 1315) has detailed for Sweden (more for Stockholm than for the rural areas of the country) the existence of conflicts between doctors and midwives. However, in comparison with “the recurrent theme” of those conflicts in the American and British experiences, he has concluded that “the studies addressing the professionalisation of Swedish midwives […] have found few conflicts between doctors and midwives. There was a gender division in the professionalisation process; however, since doctors and midwives were disseminators of the same discourse and worked toward the same goal, they complemented rather that competed against each other, unlike in the US urban setting.”
Similarly, Ciara Breathnach (2015) has found evidence of conflicts between doctors, midwives and indeed clergy in late Victorian Ireland.
This was noticed, for instance, by the distinguished Irish obstetrician and gynaecologist, Henry Jellett (1872-1948) who wrote extensively on the subjects of midwifery and gynaecology. See Jellett (1929) and also the entry for Henry Jellett in James McGuire and James Quinn (eds.), Dictionary of Irish biography. From the earliest times to the year 2002 (Cambridge, 2009).
For Ireland see Whyte’s (1980) classic study. More recent works include Ferriter (2004), Fuller (2004), Breathnach (2015). For Italy, we limit ourselves to some recent studies, Dau Novelli (2003), Gribaldo et al. (2009), Dalla Zuanna (2011).
Filippini (2020) reports also some data for Great Britain in the same period. Of the 131 caesarians performed in Great Britain up to 1880, 108 women had died according to the figures collected by Thomas Radford (Pundel, 1969, 201). On the history of the caesarian section, and the risks for the mothers, see also Mazzarello (2015).
See Betta (2006) for more details.