132 THE ROMAN EMPIRE
Sources
The evidence for imperial society is limited in quantity and quality. While
these defi ciencies should not be allowed to determine what historical
questions are asked, they do circumscribe the fi eld of questions to which
convincing answers can be given.
For the social historian, the two principal types of evidence for the Roman
empire (outside Egypt) are literature and inscriptions. Many kinds of
evidence on which historians of other periods rely never existed under the
Principate or have not survived. No systematic, self- conscious description or
analysis of imperial society and its constituent elements was written. Though
Romans did use written documents to establish legal relationships and
obligations, these have not survived in quantity; nor has reliable statistical
evidence, such as population fi gures.^2
The period did produce a substantial corpus of literary works in many
genres, ranging from history and biography to letters, legal treatises, satirical
poetry and prose fi ction laced with fantasy. The diversity offers some
safeguard against generalizations based on any one genre. For all the variety,
however, the literature was written by a tiny fraction of the population; the
authors were uniformly men of the leisured elite, and their works convey the
perceptions of the upper strata of society.
The hundreds of thousands of inscriptions form the largest body of
evidence from the Principate. But only a handful are long enough to provide
much insight into social relations, and the great majority are brief, formulaic
funerary or career inscriptions. They do not constitute a genuine sample of
the millions originally erected, nor did those millions evenly represent the
populations of the empire over space, time or social group.^3 This is partly
because the standard epitaphs and honorifi c, career inscriptions are
essentially an artefact of Romanization, which did not affect all areas of the
empire equally. Nevertheless, the epigraphic evidence broadens the historian’s
vision insofar as it issued from groups outside the imperial elite. Most
dedications were set up for modest people who did not enjoy privileged
rank. The poor, of course, are not represented even in this record.
Finally, the literary and epigraphic sources share the limitation of being
highly sporadic in nature. The lack of a series of comparable literary works,
or of a representative sample of inscriptions whose distribution over time
can be taken as signifi cant, makes the identifi cation and explanation of
trends in Roman society very diffi cult.
Class analysis
In recent years, the problem of analysing persisting social inequalities has
been presented in terms of the need to characterize or label the divisions in
Roman society. Should these divisions be identifi ed as status distinctions in