152 THE ROMAN EMPIRE
Romans... pushed things to the limit of logic’ so that the principles stand
out ‘in sociologically misleading clarity’.^4
It is important to be clear about the limits of the evidence. Roman private
law, the largest body of evidence for the family, is obviously indispensible,
but legal rules are not a direct refl ection of current practice: they could be
modifi ed through written agreements or disregarded. Though written
documents were regularly used by the propertied class, too few have survived
to give the social historian a sense of what was typical. The literary sources
provide a corrective with statements revealing common expectations
regarding family behaviour, but they are brief and written exclusively by
upper- class males. The elite bias can be overcome, to a limited degree, by
surveying the tens of thousands of funerary inscriptions of more modest
Romans, which offer some important information about demographic
variables and patterns of marriage.
Defi nitions
An understanding of the Roman family should begin with the linguistic
categories of the Romans. The obvious Latin words for ‘family’ are familia
and domus (‘household’), but neither has the semantic range or emphasis of
‘family’ as it is used today with a standard meaning of ‘father, mother and
children’.^5 The jurist Ulpian ( Digest 50.16.195) described the various
meanings of familia , beginning with the distinction between familia as res
and as persons. When used for persons, familia could indicate (1) all those
under the father’s power ( patria potestas ) including the wife (in a manus
marriage), children, the sons’ children, and adopted children; or, more
broadly (2) all agnates (that is, those related through male blood who derive
from the same house, including brothers, their children and their unmarried
sisters, but not the sisters’ children); (3) all related through males to a
common ancestor (in other words, the gens or clan, which shared a common
nomen ); (4) the slave staff of a house, farm or other organization.
The fi rst defi nition and the legal content of patria potestas have been
largely responsible for the traditional image of the Roman family: a
patriarchal household ruled by the paterfamilias (the oldest living male
ascendant) and including his wife, his sons and unmarried daughters, and
his sons’ children. For a number of reasons, it will be suggested below, this
image does not correspond to the reality very well, and, indeed, this
conclusion is supported by linguistic usage: familia in Ulpian’s fi rst sense
simply does not appear in the literature of the late Republic and early
empire. When Cicero in On Duties , for instance, discussed family obligations,
he referred to wife, children and household ( domus ), but never to familia
defi ned in this way. Under this defi nition most wives of the classical period
were not in their husband’s familia because they were not married in a
fashion to bring them under the authority ( manus ) of their husbands, and a