240 paola sacchi and pier paolo viazzo
composition and to estimate nuptiality, mortality, and fertility levels. This utterly
exceptional documentation reveals that 21% of the households recorded in Roman
Egypt consisted of more than one married couple, while another 15% contained
unmarried kin, often an elderly aunt or uncle. Over one-third of all households were
therefore either multiple or extended, to use Laslett’s typology. Even more signifi-
cantly, at least two-thirds of the people in the sample lived in such “complex” house-
holds. Hübner’s first move is to use these figures to cast serious doubts on the
consensus existing among Roman historians that the nuclear family predominated in
the Roman west, or indeed in the ancient Mediterranean. This consensus originates
from a large-scale epigraphic study by Saller and Shaw (1984), which proved deci-
sive, along with their computer-generated simulations which indicated that many
Romans would not have had a living father by the time they entered their 20s, in
undermining “the image of the three-generational agnatic household with a tyrannical
patriarch at its head” (Dixon, 1992: 5–6). Hübner meticulously exposes the weak-
nesses of the epigraphic evidence and argues that on the basis of our present source
material it is impossible to make any conclusive statements about household struc-
tures in the Roman west. However, her second move is to contend that a reasonable
idea of what they were like in the Roman world, east and west, can be given by
“comparative studies for later periods, from late medieval times up to the nineteenth
century” (Hübner, 2011: 90). To this effect, she produces data showing that the
proportion of “complex” households in the Roman Egypt sample was very close to
those found in nineteenth-century Egypt—thus suggesting a very long-term conti-
nuity in the same area—and in other historical populations inhabiting various corners
of the Mediterranean basin, from fifteenth-century Tuscany to Ottoman Anatolia
and early-modern southern France—thus suggesting that the Roman Egyptian case
should be seen as an instance of the “Mediterranean family” as originally outlined by
Smith (Hübner, 2011: 77–80).
Hübner’s argument is remarkable both because of her defense of “methods of
comparing patterns of household formation and composition that transcend bounda-
ries of centuries and societies” (2011: 83) and because of her intimation that complex
households prevailed in the Mediterranean from antiquity to the modern age and
were ultimately one of its distinguishing features. Her work confirms that a compara-
tive approach attentive to functional interrelations, to the importance of demographic
constraints and, not least, to cultural continuities and contiguities has still much to
commend itself. At the same time, it also lays bare the danger not only of overlooking
regional variations but also of freezing time into an histoire immobile which makes
wide-ranging hypotheses stressing temporal stabilities quite vulnerable to falsifica-
tions. In fact, a number of studies from Owen Hughes’s classic examination of dowry
in Mediterranean Europe (1978) to the book by Todd (2011) on the origin of family
systems alert us that customs could fade and then reappear, and that family structures,
too, could change rapidly and were sometimes replaced by other structures bearing
morphological resemblances and yet the product of historical discontinuities.
The risk of falling into the trap of inferring continuity when similar family forms,
or marriage patterns, are documented in the same place at two different points in
time lurks in Hübner’s emphasis on the similarities between household structures in
the same place—Egypt—in Roman times and in the nineteenth century. Such a risk
can, however, be easily exorcized especially for Egypt thanks to a rich array of sources