74 THE ROMAN EMPIRE
Methodologies and debates
In the revised edition of his Ancient Economy , Moses Finley challenged
conventional historiographical methods and assumptions in typically
uncompromising fashion: ‘Any analysis of the ancient economy that pretends
to be more than a mere antiquarian listing of discrete data has perforce to
employ models (Weber’s ideal types).’ In his recent Ancient History: Evidence
and Models this message is reinforced with a detailed analysis of the
weaknesses of the ancient evidence. ‘The ability of the ancients to invent and
their capacity to believe are persistently underestimated.... The insuffi ciency
of primary literary sources is a continuing curse.’ Written documents
‘constitute a random selection in both time and place, and they often lack a
meaningful context.’ Archaeology cannot uncover ‘economic structure’ or
the ‘social mode of production’, and is too often used to provide support for
the literary tradition. So we come full circle. Finley’s critique of the ancient
sources is not to be mistaken for blanket condemnation, while his models
are by defi nition simplifi ed and capable of refi nement and emendation,
largely by reference to the primary evidence. But the contrasts between
existing historical methodologies are as striking, and the fragility of much
conventional historiography is as genuine, as he has indicated.^2
We begin by considering an argument relating to the nature of the Roman
economy that has been very infl uential despite the patent weakness of its
empirical base. A.H.M. Jones suggested that the tax revenue derived from
agriculture ‘was something like twenty times’ that derived from trade and
industry in the late Roman empire, and went on to assert a rough
correspondence between this apportionment and the economic structure of
the empire.^3
Jones’ calculation, according to which the contribution of trade and
industry to the imperial revenues and to the overall wealth of the economy
amounted to around 5 per cent, was arrived at by a comparison of the
payments made by the northern Mesopotamian city of Edessa to the collatio
lustralis, often called a trade tax, in the late fi fth century, with the returns in
land tax from the Egyptian nomes of Heracleopolis and Oxyrhynchos in the
sixth century. Apart from doubts about the legitimacy of this comparison
(tentatively aired by Jones himself), one might question the value of an
equation that includes among revenues from ‘trade and industry’ levies on
usurers and prostitutes, who were subject to the collatio, while excluding
tolls, sales taxes and above all customs dues, that were collected separately.
There are more fundamental objections. Any estimate of the value to the
state of agriculture, on the one hand, and trade and manufacturing, on the
other, that is based on their relative contributions to tax revenues, will be
worthless unless we can compare, fi rst, the rate of taxation, and secondly,
the extent of state ownership, and hence the scope of a taxable private
sector, in each section of the economy. Needless to say, such information is
lacking, whether for the late or the early empire.