The Savage Wars of Peace: England, Japan and the Malthusian Trap.

Brown, Philip C.

The Savage Wars of Peace: England, Japan and the Malthusian Trap, by Alan Macfarlane. Oxford and Malden, Massachusetts, Blackwell Publishers, 1997. xvii, 427 pp. $66.95 (3 Maps, 1 Figure).

Of the problems at the center of the "new social history" and "cliometrics," perhaps none has generated such a broad flurry of activity among a relatively wide range of scholars as those associated with population dynamics in the early modern world. Whether qualitatively oriented studies of related areas -- sex, disease and epidemics, for example -- or sophisticated statistical analyses based on family reconstitution, the impact of these studies continues today. In institutional form, that activity is represented by the continued work of the Cambridge Group for the History of Population and Social Structure and such projects as the international comparative study of demographic change in a half-dozen European and Asian countries headed by Hayami Akira.

Initially grounded in number-crunching of fertility and mortality rates, the subject expanded to include elements of the history medicine and disease, diet and eating habits, and other aspects of human ecology that help us understand the processes by which population patterns began to change in key areas of the world. In addition, new intellectual tools were exploited, especially from anthropology.

In a number of ways, The Savage Wars of Peace combines these approaches. Alan Macfarlane, professor of Social Anthropology at King's College, has deeply embedded himself in the study of demographic change for more than twenty years, and has taken an especially interesting intellectual journey, conducting research in British history (in which he was trained) but also in Nepal, for comparative purposes. En route he has touched base at the Cambridge Group for the History of Population and Social Structure. Here he extends his comparative reach by undertaking an extensive comparison with Japan.

Savage Wars is simultaneously a tremendously exciting and frustrating book. It presents us with an extraordinarily rich set of hypotheses, subtly woven together, that seek to explain how some nations broke into economic growth without tragic demographic consequences. The breadth of considerations he undertakes no doubt owes a great deal to the social anthropological company he has long kept. Yet the evidentiary basis for some of Macfarlane's arguments strike me, as a Japan specialist, as inadequate and problematic.

The point of departure for Macfarlane is a demographic transition for early modern England and a rejection of such a phenomenon for Japan at the same time, relatively low fertility and mortality already having been established (p. 36). At the heart of Savage Wars lies the question of what makes fertility and mortality patterns in England unusual compared to other early modern societies? Japan serves as something of a laboratory through which Macfarlane can test ideas that originated in his research on England and Nepal. How did these two societies avoid the two Malthusian traps of first, high mortality, and second (when the first was "solved") avoiding excessive population growth? About 45 per cent of the material is devoted respectively to England and Japan, with the balance consumed by comparisons to Europe, China, and other regions.

Part I, "The Trap" begins with a discussion of the Malthusian dilemma and then discusses general overview of population dynamics in early modern England and Japan. Part II, "Wars of Peace," contains four chapters that examine the interaction of the natural environment, culture, and labour; the impact of wars; trends in the severity and frequency of famines; and diet. "In the Body," Part III, deals with changes in fluids-born diseases such as dysentery, drinking habits, and human waste disposal. Part IV, "On the Body," treats vector-born disease, the public environment, housing, clothing, bathing, and changes in thinking about dirt and cleanliness. Part V, "In the Air," consists of a single chapter on air-borne diseases. Part VI, "In the Womb," explores factors affecting fertility, the biology of reproduction and contraception, abortion and infanticide, and heirship strategies. Conclusions are summarized in Part VII, "Outcome."

 Through twenty-one chapters, Macfarlane concludes that there was no single
 demographic transition from high mortality and fertility up to the
 nineteenth century and then low mortality and fertility after that. It
 would appear instead that by the fifteenth century, at least, the birth and
 death rates had stabilized at a lower level than is normally found in
 agrarian societies (p. 371).

Macfarlane resorts to a complex, interactive set of factors rather than monocausation or even limited-number-of-causes. In the context of disease, for example, he stresses that while the viruses and bacteria may have changed with time, it is more important to examine interactions among diseases, and between diseases and changes in population density and the living environment. A variety of changes in daily habits -- drinking tea or beer, changing from wool to cotton clothing, increasing emphasis on cleanliness in the home and of the body -- that affected the spread of disease took place long before people understood the direct medical links between these habits and disease. Eating habits are also part of the picture, but Macfarlane stresses the import of drink in improving health. In both England and Japan, improvements in nutrition and availability of food meant that they did not suffer from periodic famine.

Stated in this manner, Macfarlane would seem to be presenting us with a variety of pairs of cause-and-effect explanations, but this impression is deceptive. He stresses instead the variety of multi-link causal chains in which something is done for one reason but has unintended beneficial outcomes. The Japanese construction of flexible, light houses lacking load-bearing walls was designed as a protection from frequent earthquakes, but the resulting design allowing the free flow of fresh air which was healthful.

He is also careful to stress that no one cause or set of causes is clearly "necessary and sufficient" (p. 379) to explain population behaviour that supports sustained economic growth. Variable multi-causal patterns may underlie the different experiences of individual societies that escape the Malthusian trap. Macfarlane suggests that the actual processes of interaction are close to evolutionary biological explanations of "blind Variation and Selective Retention" (p. 385). To understand these long and complex chains of interaction, scholars must adopt a holistic approach to the study of factors that condition a population's condition and interaction with economic forces, lie concludes that changes in environment, eating habits, and the like need not be major in order to have a significant impact and that even the beneficial impacts of changes were not an unadulterated blessing. Both of these considerations complicate the process of analysis.

With all of this as a background, Macfarlane concludes that a central, necessary (but not sufficient) cause of the unusual patterns found in Japan and England was the fact that both were islands. This limited warfare by creating a barrier to armed assault. Reduced defense costs and effective political engineering that reduced or eliminated domestic turmoil resulted in lower tax burdens as well as reduced mortality. In addition, Macfarlane notes that it could also hinder invasions of bacteria. In the end, he places a great deal of emphasis on the role of serendipity.

Macfarlane mines a wide variety of sources. He provides extensive and fascinating scientific background to many aspects of his story. Details of daily life habits drawn from extensively researched diaries and contemporary studies make reading enjoyable and fascinating.

Despite my fascination with his approach and much of the material, I found myself puzzled and unsatisfied with the evidence garnered. Although this reaction was true even for some of the European material, it was especially true for the treatment of the Japanese data. Let me take up several examples.

Macfarlane, using early seventeenth-century estimates by Hayami Akira and then eighteenth-century data compiled by the Shogun, suggests rapid population growth from ten millions in the late sixteenth century to twenty-six millions by the end of the seventeenth, after which population levelled off at about twenty-six to twenty-eight millions. Clearly some kind of transition has taken place if we are to believe these data from the eighteenth to nineteenth centuries.

How do we explain this? Macfarlane's emphasis is first and foremost on control of mortality to account for population increase (p. 35; see below for further comment) and an intentional limitation of family size in order to stabilize the population (pp. 304-5). The data presented on infanticide are ambiguous, and Hayami's estimate of twenty per cent infant deaths within six months after birth (p. 34) hints at infant mortality that is; probably comparable to that of contemporary England.

Yet the picture presented here on infant mortality, and in regard to infanticide (which Macfarlane acknowledges as widespread) is not one with which I am comfortable due to significant problems in data that struck me as elided. There is little evidentiary means to get a handle on either phenomenon. Infanticide and infant mortality for children of less than one year is simply not recorded in the most common sources of demographic data in Japan, the Records of Religious Affiliation (shumon aratame cho). That means that any infant who died before formal registration at the New Year is left out of our data base. Hayami's estimate is a punt, and while he is an extraordinary scholar, readers deserve a fuller discussion of problems of evidence, not just an appeal to Hayami's authority if we are to accept this figure, especially since. Hayami's figure is even lower than that for infant mortality in the 1920s which suggests that it is rather generous.(1)

Even if we accept this figure, and even if we accept Macfarlane's estimates of voluntary family limitation through cessation of intercourse with one's spouse and infanticide, one wonders if infanticide was merely transferred from family to prostitutes, off the records that historians use for demographic analysis because men sought gratification elsewhere.(2) Demonstrating the impact of conscious family limitation as opposed to conquest of infant mortality is central to determining that mortality was relatively controlled.

Macfarlane also places a heavy emphasis on population data compiled by the Shogunate from the eighteenth to nineteenth centuries to bolster his argument that Japan's population growth was stable from the eighteenth century (p. 32). Yet the quality of the data is quite suspect. No records describe how these data were compiled. Consequently, we do not know whether they included samurai, urban populations (which some other compilations of this sort clearly leave out), what age groups they included (Kaga domain's village population data count only those aged thirteen or older, not age one or older as in the Records of Religious Affiliation that most demographers use), or whether standards of data reporting are consistent across space and time. The manner in which the Shogun chose to aggregate these data is indicative of an interest in image over accuracy: Data were organized by province and county -- units that had no administrative significance at the time but harked back to the ideal administrative divisions created by the Imperial government in the seventh century.

There are basic factual errors. On mortality-related matters, Macfarlane indicates (p. 71) that there are only three famines of import in the early modern era and none occurs before 1732. This ignores a major famine in the first half of the seventeenth century, the Kan'ei famine, and at least one major regional famine in the late seventeenth century during the Joko era. While we do not have demographic data for these famines, we do have abundant qualitative data. Furthermore, record-keeping practices discussed below may simply make it more difficult to assess quickly the extent and frequency of regional incidents.

In discussion of seventeenth-century famines, as in a number of other critical cases, Macfarlane relies on the silence of English language sources as the basis for his conclusions. I readily admit that silence can speak volumes, but we must consider very carefully what subject a silence addresses. Here, I'm afraid that the silence speaks far more loudly about the lacunae of Western (and sometimes, Japanese) scholarship than anything else. This scholarship has stressed the positive side of Japan's pre-modern history -- economic development, institutional development, cultural development -- and not the challenges faced by ordinary farmers and urban residents. For most Western historians, famines do not have any relationship to the problems that they are interested in researching, creating significant gaps for anyone interested in long-term demographic trends.

In a related vein, different institutional structures have an impact on the availability of data. Reports of incidence of regional, as opposed to "national" famines, represent a case in point. Macfarlane contrasts China with Japan to indicate that regional famines were absent or not particularly significant factors in Japanese mortality. Based on my own work in local Japanese materials, I conclude that part of this impression is likely an artifact of governmental structures in the two countries during the early modern era. China had a government that truly functioned as a central administration, collecting data on local conditions nation-wide. Japan's shogunal administration functioned much more as a chief-among-equals. While it made a token effort to collect certain kinds of national data -- productive value (kokudaka) and rough population, for example -- it made no regular effort to systematically inventory local conditions such as crop shortfalls. To find regional data on famine incidence one must look at local domain sources such as records of villages petitioning for reduced taxes after poor harvests. These exist in abundance for areas I have researched in northwestern Japan (the areas of modern Ishikawa, Toyama, and Niigata prefectures) and they suggest persistent difficulties.

Similarly, in his treatment of the destructiveness of war, Macfarlane argues that medieval warfare was not particularly destructive and notes the paucity of wars mentioned in two major surveys (pp. 57-58). Yet neither of these sources deal with the wars of the late sixteenth century, by which time major combatants were fielding armies of many tens of thousands. War sufficiently disrupted agriculture that most scholars estimate that much of the "new" seventeenth-century land reclamation that supported population increase was, in fact, re-cultivation of fields abandoned during the sixteenth-century wars. Certainly one reason for seventeenth century population growth was the establishment of peace.

One final point regarding mortality. Records compiled by temples of the deaths of parishioners (kakocho) provide a less than sanguine picture of mortality in Japanese villages. Indeed, they suggest substantial impact of disease among other factors.(3) Macfarlane treats the one significant English study that employs these materials as showing higher mortality due to regional backwardness. That may be part of the story, but the comprehensiveness of the death registers may simply be giving us more reliable data than the Records of Religious affiliation which served a different purpose.

In sum, while Macfarlane never claims Japan underwent a transition to a modern mortality regime, I am not convinced by the data he presents in support of the contention that mortality was substantially reduced during this era.

In other regards, interpretation of data is suspect. Macfarlane posits that a "dominance" of a market economy made it possible to alleviate famine in one part of the country with food from another (p. 84), but the growth of a national market in the seventeenth century reflected the shipment of tax rice -- the bulk of the crop -- to Osaka for conversion into cash. Once in urban areas, rice did not readily flow back to the countryside, and inter-regional relief during famine was far more the exception than the role for most of the early modern period. Barriers to the free flow of goods were substantial, and most domains appear to have thought about economics in mercantilist terms, not free trade. The "national market" is not a good explanation for a limited impact of famine.

Macfarlane notes (p. 70) that Japan is heavily dependent on one crop, rice, and that should make it especially prone to famine. However, scholars agree that rice was not a mainstay of diet for the vast majority of the population who lived in villages and rice output should not be treated as a critical problem in shortfalls of food for most of the population. Coarse grains and tubers were more central through the nineteenth century. If there is a trend that reduced the severity of famines over time (still a debatable proposition), it lies in the growth of double and triple cropping that permitted recovery in basic foodstuffs after one crop failed.

While famines may not have produced streams of refugees to the large cities as in Europe (p. 74), this does not necessarily support the argument of limited famine impact. In Japan' a standard pattern in hard times was for town folk to go back to the countryside. That, after all, is where the food was grown, and mountain areas provided edible roots and grasses even when crops failed.

Furthermore, when villagers wanted to raid storehouses, they did so by attacking local merchants in sometimes quite violent attacks called ikki. On these, there has been a small explosion of English-language study since the 1980s, but Macfarlane is unaware of them.(4)

In other cases, the data Macfarlane presents are simply anachronistic. Late seventeenth- and early eighteenth-century farmers' writings indicate that the key fertilizers in what must have been the transition period (seventeenth century) were not the kind of commercial fertilizers Macfarlane cites (p. 155) in support of an agricultural revolution of sorts. The source cited, King, Farmers of Forty Centuries, is an early twentieth-century source that does not clearly distinguish how much had changed in agriculture between the mid-nineteenth century when Western influences began to transform Japan and the time that King was writing.(5) In a less critical context Macfarlane notes that one way of signalling that a young woman was available for marriage was through hair styles. Unfortunately, the respondent who provides this information (p. 309, note 37) was born in 1918. Furthermore, if one looks at the styles depicted in the source cited, it becomes clear that these are not everyday hair styles, but those for special occasions. Citing this kind of anachronistic material is not necessary, since most early modern rural communities were small enough that everyone knew who was available for marriage without the kind of telegraphing depicted here. Unfortunately, the use of nineteenth century data to reflect on seventeenth century changes is very common. See Chapter 11, "Public Environs, Streets, Fields and Markets," for example. While long-term structural factors merit consideration, Macfarlane should have been more sensitive to nineteenth century changes.

Macfarlane's logic is occasionally tortured and based on stereotype. He notes, by way of partial explanation for population growth in the late nineteenth century, that the Meiji government forbade abortion and infanticide. He argues that "a population as law-abiding as the Japanese" (p. 341) would have followed these laws, leading to a sharp drop in the use of these forms of population limitation. Even if we accept this as true for Meiji, it begs the question of why "law-abiding Japanese" in the seventeenth to mid-nineteenth centuries would have ignored frequently issued injunctions against infanticide! Furthermore, Macfarlane presents twentieth-century evidence that contradicts this impression of "law-abiding" behaviour (p. 348).

Issues such as those I have outlined raise questions about how to undertake a comparative history project. Such efforts are, I believe, very important and potentially very useful. They can do a great deal to help us break out of parochial intellectual traps and expand the range of possibilities we examine in a given regional-national context. I find Macfarlane's adventuresomeness and ambition laudable, but I wonder if it is possible -- short of team research -- to test so many diverse hypotheses when examining a society or societies with which one is not familiar. Macfarlane consulted scholars in Japan, but there was no close collaboration. Discussions with Japanese demographers, useful for some facets or research, are inadequate for others. Issues such as the impact of different institutional structures on the generation of data noted above do not generally enter the consciousness of demographic historians since the data fall outside those of their normal purview. To me, problems of this sort and others such as those raised above, suggest the need for close consultations with historians broadly familiar with the regions being compared, if not the construction of a formal research team.

I am grateful to Macfarlane for providing new perspectives for me and other historians to consider. However, :given the problems that result from the failure to work closely with Japanese historians, I find it difficult to have confidence in Macfarlane's conclusions, no matter how stimulating I find his hypotheses.

Ohio State University

(1) The data for 1926 are cited in Laurel L. Cornell, "Infanticide in Early Modern Japan? Demography, Culture, and Population Growth," Journal of Asian Studies, 55:1 (1996), 29. For those unfamiliar with Japanese demographic history, this article provides useful summaries of problems of data and major interpretations of trends in the early modern era.

(2) The evidence for "stopping," drawn from sources substantially available to Macfarlane prior to publication are summarized by Cornell, pp. 34-39.

(3) Ann B. Janetta and Samuel H. Preston, "Famine Mortality in Nineteenth-Century Japan: The Evidence from a Temple Death Register," Population Studies 45:3 (1991), 417-36.

(4) Most pertinent is the work of James White, especially The Demography of Sociopolitical Conflict in Japan, 1721-1846, Japan Research Monograph, 12. Berkeley Institute of East Asian Studies, Berkeley, CA, 1992.

(5) For a discussion inkling of the degree of change in agricultural output, see "Rumbles in the Ricefields: Professor Nakamura vs. The Official Statistics," Journal of Asian Studies 27:2 (Feb 1968), 347-60, in which even the most conservative estimates (Nakamura's) show quite impressive transformation.