Light on a rich country
- The Population History of England 1541-1871: A Reconstruction by E.A. Wrigley and R.S. Schofield
Edward Arnold, 779 pp, £45.00, October 1981, ISBN 0 7131 6264 3
The title of this book means what it says: it is about England, not England and Wales. The exclusion of the Celtic fringe can be explained by the very real difficulties which arise for some forms of historical reconstruction from the narrow range of Welsh surnames and the weakness of the Established Church in Wales. The work is based on the extraction of figures from parish registers, the Census, once it was established, ‘family reconstitution’ carried out for a dozen parishes, a central group of experts using sophisticated numerical methods in Cambridge and the strengths of computerisation. So we have chapters in which the figures of events and base population are built up, and then chapters in which social and economic conclusions, based on these figures, are worked out. The figures are tucked away into an enormous structure of appendices, 16 in number, amounting to over a third of the whole book, but everything hinges on them, so the book has to be taken as a whole.
The unitary nature of the research means that a review must discuss how the figures are derived. Before the first Census of 1801 there exist partial counts of population and estimates for various dates. Before the start of civil registration of births, deaths and marriages in 1838 there exist Church of England parish registers of the accompanying ecclesiastical events, baptisms, burials and marriages, with gaps when the register was lost, the parish clerk ill, the benefice unfilled, the village upset by civil war. Yet here in Appendix 2 we have page after page of figures for these vital events on a national scale, and also figures for the size of the base population, year by year, in which they occurred. How has one type of information been turned into the other?
The Cambridge Group for the History of Population has been funded as a specialised research unit since the 1960s – first of all privately, later by the Social Science Research Council. Its members have been partly freed from other academic burdens to carry on research and maintain one of the liveliest centres of debate in the British academic world. The debate has been both within the Group’s offices in Cambridge and with the general public, for the members of the Group have looked to private individuals to examine and analyse local parish registers, and have founded and sustained the journal Local Population Studies to publish some of the results. These volunteer researchers collected, month by month, material from 530 parishes: examination by the Group reduced this number for its own use to 404, all of which covered the whole period of 1661-1811 with only minor gaps, and it is their monthly and annual totals of baptisms, burials and marriages which form the basic material of the book. They are enriched by the information from the ‘family reconstitution’ parishes.
Even the hand-picked 404 parishes had deficiencies: these were identified by rules specially formulated, and then the entries which were considered to be missing were supplied on the basis of the identifiable trends which the gaps had disrupted. The authors are careful to refer to this process as ‘replacing’ – rather than ‘adding’ events to the registers. The annual totals then underwent a further restoration process in adjustment for under-registration. There was in the 16th century no organised dissent which refused the services of the established church, and babies were brought to church for baptism immediately after birth, so that in the early part of the study registration can be considered as equivalent to the events recorded. By the 19th century, dissent, separate burial grounds and chapels, laxity and secularism had all worked on this weak link in the record, and registration was increasingly remote from the actual number of events. So the annual totals had to be enlarged on a changing schedule, to fill in the unregistered items.
After this, the Group had to consider to what degree the 404 could be taken as a typical sample of the ten thousand or so parishes in England. They came from pastoral as well as grain-growing areas, from every county except Cornwall and Westmoreland, and in all sizes. Still, they tended to be southerly in location, and the large communities were overrepresented through an initial bias in their favour, based on the idea that such large parishes would be relatively free of unrepresentative variation. Large parishes were found, in fact, to have grown faster than other parishes, so the blocks of registration figures had to be weighted to discount this, and weighted in different proportions for the different series over varying periods.
After all these adjustments there emerge monthly and annual figures of births, deaths and marriages. A special gap that had still to be made up is the Great Wen, London, for it had been an initial, if mistaken impression that parish material for London could not be used in the same way. Another special problem was created by the years between 1811 and 1838, when parish registration was deterior-rating fast. Figures for this period are derived with special attention to the mortality of women, since it can be assumed that they did not share in the processes which involved men in the wars. In this period, also, the Princeton Model Life Tables, a volume which would earn Charles Lamb’s description of a ‘thing’ in book’s clothing, are used to estimate vital events. The rapidly changing mortality pattern of the period appears to lie within the systems covered by the Princeton North tables. Another problem patch is the 17th-century Interregnum, for replacement of ecclesiastical by secular registration created its own problems of under-recording. Some of the adjustments made are fairly arbitrary: in 17th-century London, plague deaths, worked out from the Bills of Mortality, are added in in surges to August mortality – a pattern not entirely borne out by what we know of the timing of epidemics. After all these problems have received a numerical answer, the series of vital events is made up for nearly three and a half centuries.
But to what sort of population do these items belong? A small population with high mortality and fertility could produce figures very similar to those of a larger but more tranquil system. Until the size of the base population is known, vital rates cannot be extracted, let alone compared over time for the perception of long-term change. The solution to this problem lay with the computer. The technique devised is called ‘back-projection’, working from the age and sex distributions recorded in the 1871 Census. The researcher works backwards in five-year steps from 1871, adding in for each span of time a group in the oldest age band, and calculating the other age bands by their survival chances to the period ahead, and the youngest age groups by birth and death information. The first check comes when the invented old-age group is taken back to the period of their birth. If the figures do not tally, the whole process has to be done again with a different figure added for the oldest age group. Somewhere along the line, too, allowance has to be made for people born in England but dying elsewhere, for emigration and lives lost at sea or war. The imperialisms of several centuries have to be half guessed, half calculated.
Back-projection works by repetition and correction. If wrong numbers are fed in, the resulting error builds up exponentially until it should be too large to be missed, and then adjustments have to be made, and another run set out. It is an example of the method of solving a common type of mathematical puzzle. Put in a figure and work through its consequences until they either give a consistent and believable answer or become self-contradictory. A sample of this type is the numerical solution of the multiplication sum, BALL × BAT = TEAPOT. It is easy to see that L must equal 1, and that B can be only 2 or 3. To get any further one has to plump for one or other figure for B and go ahead. Of course, in this sum, which is stolen from a book of puzzles by Hubert Phillips, we have assurance that there is one and one only correct answer. In historical demography there is no such assurance, and no list of answers at the back. But there are constricting ranges of the possible, some set by the ecclesiastical censuses of the 17th century, some by previous studies and estimates, notably the work of Gregory King and Rickman, as well as the opinions of modern scholars. The figures for the total population of England at various dates from back-projection are bracketed by these estimates and partial counts. The structure of the argument is at its weakest for the 16th century, for then the age groups fed in can receive no correction from birth cohorts. Here it is consistency which must give confirmation.
Once a base population history is known, the type of population can be found. Birth and death rates are set down, and then the Gross Reproduction Rate is calculated. Family reconstitution supplies other figures, the proportion of the population married and the age of marriage, but this information is less secure than other types.
Like an old building, the total structure has some parts which perhaps an owner with a new and single-minded approach would have designed differently. It was probably a mistake to leave out direct work on London. It was certainly a mistake to confine the basic material to parishes with almost complete long runs of figures, and not to supplement for the period before 1661 from others with material for the first 120 years of registration. There are fewer than a hundred parishes to give a base for the period before the accession of Elizabeth I. It may also have been a mistake to have selected parishes by size, for the pattern of child and infant deaths seems to have differed according to the size of the parish. But the Cambridge Group were working with over three million figures, and it would be preposterous, and ungenerous, to argue that they should have gone out and looked for more.
Doubts are bound to be raised about whether the picture in this book, which in many ways is very surprising, is the truth, even whether we can accept the basic tables of the appendices. These doubts have to be met at the level of basic historiography. Demographic history is history about numbers: it starts by collecting them, and then adapts and interprets them. If any progress is to be made, then, once the numbers have passed carefully devised tests, they must be used as facts. Otherwise we are back in the sort of history which uses phrases like ‘a considerable amount’ or ‘a large percentage’, making no attempts to define these terms. There is an obligation on the historian to devise methods to exploit his material, instead of generalising about what it might lead to. For this same reason we must admire the resolution here shown which presents at every stage absolute numbers, even where they are clearly only roughly true: the alternative is approximations of approximation, getting more generalised and uncertain at every step.
The startling thing about this book, therefore, lies, not in the methodology, which is dictated by the material, but in the conclusions. English demographic experience is shown to be very different from that of other Western European countries. The growth or stagnation of the population was dictated mainly by changes in the marriage pattern, and only to a small degree by changes in mortality. Crisis mortality, in particular, had little part to play after the 1550s. Population was buoyant in the 16th century, with high fertility and, normally, low mortality, so that the total rose from some two and three-quarter million in 1541 to 4.1 million in 1601 and 5.13 million in 1656. But the age of marriage was rising, and also the percentage not marrying. The age of marriage reached a plateau level in the mid-17th century and stayed there into the mid-18th century. During this century of restricted marriage much of the expansive element in the population went in emigration. Only after 1750 did the home population start to rise again, and to rise rapidly, so that it went from 5.8 million in 1751 to 8.7 in 1801.
Much of the change in totals, the rise of population in the 16th century and the later 18th, has been the current opinion of historians for some time. What is new is the mechanism that produced it. Most historians have tended to look to a decline in mortality to explain the main periods of growth, encouraged by the fact, here confirmed in numerical terms, that mortality has been a more variable component of population than fertility, but also by the fact that the late 18th-century population growth is to be found in most of Western Europe. With parallels in growth existing between countries at very different stages of economic development, it has seemed reasonable to look to some exogenous factor, such as infectious disease, rather than to changes in social organisation or social intent to explain the main change. Wrigley and Schofield attribute only a quarter of the 18th-century expansion to mortality change, and argue that there is no logical connection between the flexibility of mortality levels over short periods and the likelihood that mortality over a long time-scale is the main instrument of growth. This is true, yet the fallacious position has more strength than they allow, for it was based on clear evidence that death rates did vary sharply, whereas the existence of fertility change has needed an enormous research project before it could be seen. Those in error were comparing what they could see with what, till now, was a remote hypothesis.
The importance of the new explanation is the light it sheds on social history. England is here revealed as a rich country, at least in its southern part, largely removed from the pressures on material resources which were the common lot of early modern Europe. This is also shown by the low level of total mortality even in ‘crisis’ years. In 1558 and 1559 the level of deaths was 124 per cent above the trend of the period. After that, surges of mortality never passed 100 per cent, and crisis years, however defined, which were very rare in the 18th century, became rarer. Even at the parish level there is no sign of great surges after the 1550s. In this, the English story contrasts with those from the rest of Europe, most notably with that of Sweden, where the main influence on population change was death, and to a lesser degree with that of France, where fertility combined with mortality as instruments of change. In an indirect way, this work adds an element of explanation to the problem of why it was in England that the Industrial Revolution took place. Clearly pre-industrial England had a relatively rich and sophisticated economy in which the market element was strong enough to cope with basic harvest fluctuations, and which developed a system of public aid capable of carrying the population through temporary crises. The authors label the English demographic system a ‘low-pressure system’, in contrast to the types which relied on Malthusian positive checks, but it had to be backed, until the mid-18th century, not only by a high-pressure economy, but also by a high pressure of social discipline.
The other part of the story, the fluctuation in fertility, reveals, indeed, a remarkable level of social organisation. Recent work by, among others, David Levine has shown the likely economic developments that led to the lowering of the age of marriage and the relaxation of sexual morals in the 18th century. These changes are attributed to the rise of ‘proto-industry’ – in particular, to the proliferation of home textile manufacture, which not only made it easy to deploy advantageously the labour of children but also provided a new range of sources of income which were not dictated by the needs and resources of a local economy. Some thirteen years ago I watched Sir John Habakkuk at a conference sketch out on the board how a drop in the average age of marriage in 1740 of two years, maintained throughout the rest of the century, would explain the whole expansion of 18th-century population in England: but everyone agreed that there was no reason to think that this had happened. At that date, when we thought about the expansion of the economy through industrial development, it was to the factory system created at the end of the century that we looked. Later attention focused on the proliferation of industry earlier, which could give a possible impetus for change, and now this book sees that the change did happen, and very much as Habakkuk suggested.
There is less clarity in the known setting for the earlier period of basic change. The authors give as a general outline: ‘rapid population growth in the 16th century provoked a sharp fall in living standards which in turn was followed by a falling away in the rate of population growth so pronounced as to produce a 30-year period during which population fell and a much longer period lasting 65 years during which population was below the peak total reached in 1656.’ But they also point out that in the 17th century the rapid growth and high mortality experience of London, and the loss of over a third of a million people by emigration, played a part in the decline of population after 1656. They argue that the early modern economy could sustain a growth in population of up to 5 per cent a year, but if growth was faster than that, then living standards fell. Unfortunately, the quantitative basis for these judgments is the Phelps Brown-Hopkins real wage series. This series has long caused doubts: at times the price index has a narrow geographic base, the wages used have sometimes even less support, and ‘wages’ was in any case not a valid concept for most of the population before the mid-18th century. To make matters worse, only last year Donald Woodward showed that the building craftsmen on whose pay the wage index relies could not be considered true wage-earners in the early part of the period because they had other sources of income. As yet, no other real wage index which spans the whole period of this book is available. The implication of the conclusions here is that the agricultural changes to which Eric Kerridge has given the title of an agricultural revolution may, in the early 17th century, have enabled an expanding population to feed itself adequately, but this is not a process which can be measured in real wage terms, given the index that We have.
It seems that once again the desire of the authors to use only material which spans several centuries, and to believe the figures they are given, has led to assertions about standards of living which, if true, are not measurable, and to over-concentration on these. It is not only income level that has to be considered by a couple thinking of marriage: it is the pigeonhole in the economy in which the potential husband hopes to fit. Men had to think in terms either of recognised jobs or of a defined mixture of activities which might support a household. And here we run into real difficulties: for the period of decline and eventual stagnation in population growth, the period of late marriage and, for many, permanent celibacy, spans the great building boom of the 17th century, the spread of the domestic revolution in household standards into the north and west of England, and a proliferation of small crafts, trades and traders. England, still largely not a wage-earning society, at least on a life-long basis, presents us with a puzzle as to the motivation for the drastic curtailment of fertility achieved by curtailment of the prospects of marriage. This book shows us on a broad and quantitative scale what Pepys’s diary shows us on a small one: the relative stickiness and indifference of the family to the issue of how, if at all, a girl heading towards the status of old maid is to be married off. The Pepys family would enter into negotiations over the marriage of his sister, but at some point of the bargaining they would draw back, while the young woman stayed at home and quarrelled with her relatives.
The later chapters of the book discuss the interrelationships of price movements with basic population rates, showing, for instance, that the effects of high prices on mortality and nuptiality, which were real enough, were never very striking and always short-lived – though, as they carefully point out, there could be sharp local rises in deaths in high price seasons, indicating devastating but local food shortages. From a detailed study of the patterns of interaction they move on to a discussion of the long-term trends in their economic setting. This chapter shows the inadequacy of the Malthusian model for early modern and modern history, since it is an argument based on the concept of a closed economic system, and all the European countries had elaborate international links. It reveals instead what they call ‘dilatory homeostatis’. Fertility appears to follow real wage movements after an interval of nearly two generations. This is a conclusion which reinforces the emphasis already laid on restraints on marriage, and on sexual activity outside marriage: ‘rules of social conduct that have sufficient strength to prevent the young from marrying until their late twenties, and that keep a substantial minority unmarried for life, are also likely to be effective in preventing less permanent unions’ is their comment on low illegitimacy levels. If we are to accept that it took some fifty years for these rules to adjust to expansion in the economy and its ability to support a larger population, then we want to know on what impulses they were eventually modified, and this means that we need to be clear about whether new types of employment became available, or whether the local economic climate became more benign. Our ancestors lived with very little in the way of privacy: details of family life and worldly resources would be known by the whole of a village community, giving a constant reminder of how close to the bone things were for some families. Young people were, by the normal age of marriage, removed from their home environments, where their parents might be expected to put pressure on them to conform to the marital expectations of an earlier age. They had become ‘servants’, labourers or apprentices living in their employers’ homes, visiting their parents very occasionally or not at all. Yet if the graphs of the parallel movement of wages and fertility are more than an accident, the experiences of an older generation were in some way brought to bear – were accepted and allowed to override sexual impulse.
The book ends with a wide-spanning survey of the structure of the intercommunication of different aspects of economy and demography at different periods. It starts with a simple model of the elements of demographic change, food prices, incomes and the demand for labour, and builds this into a more elaborate structure with positive elements, such as the impact of urbanisation on mortality, modified by ‘feed-back’, the influences on the demand for goods, which sprang from the higher real wages, on the supply of goods, and so on the provision of jobs. By the early 19th century there has ceased to be a link between mortality and nuptiality: peasant society has given way to more capitalised farming, as well as to industrial work, and for neither of these would the opportunity to marry depend on the death of a parent. Other links have weakened too, even if they have not disappeared. By the late 19th century the economic system runs as one unit, and the demographic as another, for those links which still remain are very weak. This general development is labelled ‘one of the most fundamental of all changes in the history of society’, an accumulation of economic change and social adjustment revolutionary in its impact.
This book, one of several to emerge at last from the Cambridge Group, provides a basic new look at the whole development of English society from the Reformation to the industrialised world of the mid-Victorian age. Not all will agree fully with its choice of economic evidence; many will find its calculations difficult, and some of its conclusions stretched beyond the supporting evidence. But the view it expresses of changes in society, and of the complicated relationship between social and, in particular, family structure and resources, is offered with care, quantification, moderation, and austerity in argument. This makes it fundamental reading for all historians of the early modern world. It is also a landmark in the technology of number. To anyone actively concerned in understanding how societies adjust, balance and develop over time, it must come as one of the most exciting works of the last ten years.