- Chapter Eight -
further emphasizing the power this had over the ultimate direction and character of
romanization.
Towns and villas clearly represent the visibly romanized superstructure of the new
province, set alongside the obviously imposed aspects represented by the colonies and
the administrative capital and trading centre at London. Quite how the two elements,
the conservative and the progressive, articulated remains unclear, but one might
expect that if the native elite did retain a stranglehold over existing socio-economic
and redistributive networks, they would also have sought to integrate, neutralize or
marginalize the potentially disruptive influence of the 'external' islands, whilst at the
same time seeking to demonstrate their participation in the new world order which
these very same islands represented. Indeed, during the early stages of the romanizing
process and the emergence of a distinctive new Romano-Celtic society, the elite in the
towns and in their villas (aided and abetted by traders, entrepreneurs and the army)
must have set the standards for others to follow, commencing a process whereby
romanization gradually percolated to varying degrees down the settlement hierarchy
and the social order.
This process is clearly discernible in a variety of ways in the archaeological record
of the south and east, reflecting a gradient from highly romanized to less or non-
romanized, although the full details need considerable clarification. Perhaps the
clearest such gradient is to be found in the evidence for diet (King 1984). At the top
of the scale come the military sites and highly romanized cities which have a high
percentage of pig and cattle bones by comparison with sheep (up to 90 per cent on
some sites), followed lower down the scale by the major civitas centres and the villas
(Figure 8.7); below this the balance progressively shifts in favour of sheep down the
non-villa spectrum, though the gap narrowed in the third and fourth centuries. This
seems to be clear evidence for a shift towards a romanized diet, yet it had already
been happening in the later Iron Age as part of the 'pre-romanization' of the
progressive elite; now it was being extended further down the scale as romanization
bit deeper. The same is arguably true of romanized artefacts, which progressively
became available within society. Fulford (1982) has certainly suggested that, type for
type, the most extensive range of finds (especially in terms of luxury items) is
concentrated in the cities and towns, closely followed by the villas, whilst further
down the scale the range is less extensive and the balance of functional to luxury
items more pronounced. This balance is typical of the sort of gradient which might
be expected, though there are obviously chronological trends which need detailed
elaboration. Comparable gradients seem to be recognizable for sculptural and artistic
features, not least for the different categories of 'small town' considered by Millett
(1977); this impression can be extended to other sites in the settlement hierarchy
(especially the rural sanctuaries and temple sites), where the number of pieces with a
clearly identifiable, romanized symbolism is highly suggestive of the desire to appear
romanized well down the order. The same may well be true of the adoption of
the Latin language (Mann 1971; Evans 1987). Romanization may also be measured
in other ways, including such things as the shift from timber to stone and from
circular to rectangular structures, the adoption of better building technologies, the
use of more effective storage systems and wells.
The clear suggestion in all this is that the native elite, together with incomers, set