76 THE ROMAN EMPIRE
The argument about the economic activities of the rich is mirrored by,
and closely associated with, the debate over the nature of the economic life
of cities, the seat of most men of means in the Roman empire. Again the
implication is that the issue of the character of the Roman economy at large
will be illuminated, and again it is possible to proceed by means of a
conventional methodology, the accumulation of information about the
economic life of particular cities. Here the problem, supposing one can hit
upon a relatively well- documented area, is to convince anyone that the
results have a general signifi cance. Thus, for example, it is uncontroversial
that the economy of some cities was based on trade rather than agriculture.
The question is rather whether such cities were exceptional.
This is an appropriate point to turn to a consideration of less conventional
approaches, and, in particular, the construction of explanatory models,
whether quantitative or non- quantitative. Perhaps the most familiar and
infl uential model concerns the urban economy. This is the ‘consumer city’,
adumbrated long ago by Sombart and Weber and revived and publicized by
Finley. According to this model, the ancient city was primarily a centre of
consumption, in contrast with the medieval city, which was primarily a
centre of production. By a consumption city is meant ‘one which pays for its
maintenance... not with its own products, because it does not need to. It
derives its maintenance rather on the basis of a legal claim such as taxes or
rents, without having to deliver return values’.^9
The essential power and truth of the consumer city model can be admitted;
so can its role of confi rming the supremacy of agriculture in the economy.
The city was both the base of the major landowners, who were also the
wealthiest residents, and the centre and focus of their expenditures, which
were funded in large part by their rural investments. Whether the model
reinforces the minimalist view of the role of trade and manufacturing is
another matter. The model, any model, is not a statement about reality,
which is much more complex and problematic. The minimalist view is such
a statement, and must be tested along with others. In fact there is some room
for manoeuvre within the limits set by the model. Thus, for example, it can
be argued, with the aid of familiar literary sources, that city elites were not
merely holders of rural property, but also invested in urban property to a
considerable extent. Without rejecting the crucial importance to most
ancient cities of the fl ow of income from the countryside, one might want to
recognize that there were income- creating activities taking place in the
‘internal’ urban economy, activities access to which, and often control over
which, were provided by ownership of urban property.^10
This last argument is instructive for its implication that ‘model- building’
and ‘empirical’ data collection and analysis can be complementary rather
than competing methodologies. No model will carry conviction unless it can
be shown that it bears resemblance to the historical reality. With this in
mind we may consider some further arguments that carry negative
implications for the minimalist thesis.