Lebenserwartungen in Deutschland, Norwegen und Schweden im 19. und 20. Jahrhundert.

By Arthur Imhof (Berlin, Germany: Akademie Verlag, 1994. 724pp.).

This is really three books: reflections on the meaning of death, a general history of north-central Europe, and statistical documentation of mortality rates. These three products of a team at the Free University of Berlin are loosely linked by their commitment to modernization as a success story. The formal occasion for publication is to extend to the present the life tables of Lebenserwartungen in Deutschland vom 17. bis 19. Jahrhundert (reviewed in this journal, 26 [1992], 390-392). Here the spotlight is on big units, Germany in its various incarnations, over the long term. Norway and Sweden appear in supporting roles to set off the German experience.

Team leader Arthur Imhof's foreword orients readers to this ongoing enterprise. The basic finding is that life expectancy at birth doubled from 37 in 1855 to 75 in 1985. Mortality in childhood led the way down, with the decisive decline in infant deaths beginning after 1900 in Germany, somewhat later than among Scandinavians. From a pattern in which deaths peaked in infancy and were frequent at all ages thereafter, Germans have been moving towards a spiked distribution in which a large majority die as octogenarians.

Having established that great transformation, Imhof's introduction turns to his own interests, which range well beyond conventional historical demography to early modern painting and futurological speculation. The latter discussion is less informed by its own strictures regarding historical contingency than one might wish. All the same, there are useful reminders to take age seriously as a factor in history, and to think of the globe rather than Europe.

It is disconcerting then to find Imhof repeatedly referring to Europeans, Germans, men, and even the middle-aged as "we". Imhof urges on his own age-group the development of a life plan to master the additional decades vouchsafed to them in the next century. He is more modest with advice to those already retired. The cohorts born 1890-1930, having pushed back the frontiers of death, are entitled to enjoy the years they have reclaimed on their own terms, he declares. Perhaps it is uncharitable to point out that those cohorts also voted Nazi disproportionately, but that coincidence might temper the incitement of the Third World to follow Germany's modern path.

Imhof does sketch his view of why life expectancy rose, although he abjures justifying his conclusions in this work. A series of illustrations assigns pride of place to women's education, in particular their ability to read. In a graphic mixed metaphor, this factor wields a key, whose swings topple the first in a line of dominoes.

The second section of the book provides background on periods covered by the statistics and how they were calculated. Many pages are devoted to a textbook review of German history. Basing itself on multiple overlapping circular causation, in which just about everything turns out to lengthen life expectancy indirectly, the account indiscriminately includes the punctuation of Olmutz, the rarity of telephones before 1900, and the limited recruitment of university students through the 1960s. No direct evidence relates death rates to these phenomena, nor even to political persecution or border changes or medical treatment.

The briefer contribution on Sweden stimulates more concentrated insight. Here too no final answers are forthcoming, but there is a sustained effort to sharpen the questions. Anders Brandstrom and others contrast a sawmill town growing tenfold in the nineteenth century to a forested region and two nearby villages, one agrarian and one industrializing. The varying gaps between survival rates for males and females provide food for thought.

The final section comprises hundreds of graphs and tables setting out frequencies of survival by age and sex. These supplement the period life tables of the previous volume with aggregate data from Statistischen Jahrbucher since the 1880s. The shift from the earlier reliance on individual life histories in local genealogies compiled from parish registers goes largely unanalysed. This is not for lack of information. Indeed, genealogies for one region, Herrenberg in Wurttemberg, have been processed past mid-century. Raw results are presented, but comparative assessment is left to future researchers. Only they can tell whether the breaks in various curves around 1850 are artefacts of discrepant sources. To be sure, their task will be eased by the free availability of the data through the University of Koln.

The absence of a thorough source critique leaves several questions from the earlier volume hanging. The description of how family histories were selected is clearer here, but important technical points remain obscure. How, without archival research, did the team cope with such simple but obstinate difficulties as typographical errors in the genealogies? More salient, many deaths reported as occurring within the first minutes of life have been arbitrarily reclassified as stillbirths - and as such excluded from the life tables. That practice might have acquired some justification from tight comparison of ecclesiastical and civil registration, or of both with census collection. Again, until the comparison is made openly, one cannot judge the value of the adjustment.

Before such adjustments, the genealogies are on the whole complete enough and reliable enough to document general trends. In Demographic Behavior in the Past, John Knodel has used them cautiously to estimate infant and child mortality on the one hand, and the survival of spouses on the other. The Berlin project ploughs right across the intervening adolescent years of mobility to offer a uniform account of the entire lifespan. Yet it is hard to tell how well insulated the calculated rates are from the upsurge in emigration in the mid-nineteenth century. In the Swedish case, with much fuller information, emigration took 46% of the population out of observation. It is still less clear how fully the processed data reflect the experiences of minorities such as Jews, celibates, or children born out of wedlock. Researchers in those fields will have to make their own checks before employing the data reported by Imhof's team. In that respect as in others, understanding transitions to secure longevity is by no means as far along as the transitions themselves.

Ernest Benz Smith College