Reviews:Walter Scheidel's Debating Roman Demography
(Leiden, 2001) by Walter Scheidel
It is disappointing that the pre-eminent Roman demographer in America, Walter Scheidel (Debating Roman Demography, Leiden, 2001) and so many other distinguished classicists have endorsed the erroneous thesis of Richard Saller (Patriarchy, Property and Death, Cambridge 1994) and Brent Shaw – that Roman males married on average at twenty-eight and females at age nineteen, and that therefore the patria potestas was not important. As Arnold Lelis, William A. Percy, and Beert Verstraete have shown in The Age of Marriage in Ancient Rome (Lewiston, Queenston, Lampetere, 2003), the ages at first marriages of Latin speakers were normally between fifteen and nineteen for males, twelve and sixteen for females, as almost everyone had assumed before Saller began to analyze dedications on epitaphs in 1983. (See Ludwig Friedlander’s 10th edition of Darstellungen aus der Sittengeschichte Roms in der Zeit von Augustus bis zum Ausgang der Antonine, published in 4 volumes, Leipzig, 1922. Debating neglects even to cite this still unsurpassed masterpiece.)
What the epitaphs actually show is that by twenty-eight most husbands had in fact lost their fathers, who always commemorated sons as long as they could, with the result that their wives thereafter became their principal commemorators. Likewise by nineteen most wives had living children, so that their husbands, who therefore got to keep the dowry, overtook their wives’ fathers as commemorators. Scheidel has challenged anyone disputing Saller’s assertion to prove his claim, stating that the burden of proof rests on his shoulders, rather than on Saller’s. Why should this be the case – whatever modern fertility transition theory argues – when it is Saller who has made the revisionist claims against the accepted demographic interpretation?
On November 19 2005, I've received the following correspondence from Walter Scheidel:
Dear Professor Percy,
Thanks for your comments, and the excerpts from Luuk's review (I have known him for a long time but wasn't aware of the review). I have only three comments:
(1) Clustering of deaths. I didn't comment on this in my draft because the whole idea is so clearly demographically unsound. An average age of death doesn't imply any kind of clustering. Model life tables, which are not perfect but reflect the experience of actual high-mortality populations (as they are based on empirical data which they present in idealized form) show very clearly that as the probability of dying rises with increasing age and the number of survivors decreases accordingly, the actual number of deaths per age group remains fairly stable. To give just one example, in a model population with an average life expectancy at birth of 25 years, out of an initial birth cohort of 100,000, 8,699 men die between ages 45 and 55, 8,888 from ages 55 to 65, and 7,412 from ages 65 to 75. It is only afterwards (when sons who had been born when their fathers were in their 20s would already have reached their 40s and 50s) that there aren't enough survivors left to sustain these numbers and aggregate mortality drops significantly. In other words, there is no sign of a clustering of deaths around age 60 or any other age, in any known historical population. This is the reason why I argue that the Saller/Shaw reading provides a better 'fit', because it accounts for the sudden surge in spousal commemoration around age 30 for deceased men whereas your explanation doesn't. I see now that I should have made this clearer in my paper. (The other part of the 'fit' concerns the conspicuous shortage of putatively married and fatherless men who died in their 20s and were commemorated by their wives. I would consider this to be a very serious obstacle to your reading, but you do not address this issue.)
(2) Circularity. It is true that Saller and Shaw's reasoning is circular, but I am also unaware of any *independent* evidence that Roman fathers always commemorated their (married) sons.
(3) In de Ligt's review, his reference to the Augustan marriage laws (men were supposed to be married by age 25) strikes me as irrelevant because there were no real sanctions for sub-elite men who did not comply. This law was clearly aimed at the ordines, i.e. senators, knights and city-councillors, who -- as I am happy to believe -- usually married around age 20.
Best, Walter Scheidel
On Labor Day of 2005, I presumed to call Professor Scheidel. I asked him if he had heard of our book; he replied that he had, but that he had not yet had a chance to read it. I asked him to please read and review it if he had time, because I believed that it would help prevent him from making mistakes in the future. On October 24th my assistant printed out for me the following gratifying response from the professor:
- “Your book encouraged me to re-visit the whole issue in more detail, and I have now finished the first draft of a short paper on funerary commemoration and age of marriage. If you are interested, I would be happy to send you a copy for comments.”
Yesterday on November 3rd, I received the article offered to me by Professor Scheidel. As yet I have not had time to study it in detail, but at first glance I find his equivocation problematic. As you can see from the review by Professor L. de Ligt from Leyden, the acceptance of the low marriage age for males is easier than for females, in part because we have so many more examples; however, even he doubts the low marriage age for females, because he doesn’t seem to realize that most children died before six, and if the wife nursed the infant, she was unlikely to conceive again within three years, and therefore even marrying at fourteen (as I believe they did on average), they would not have living issue until nineteen. I hope that when I reply in greater detail to Professor Scheidel’s article, I hope that I can convince him to quit equivocating about the low marriage ages for males and females. - Posted 11-05-05