Discussion
Before looking in more general terms at the pattern of development in the insula, we may make a number of preliminary points. First of all, it bears repeating that the process of change which we have outlined was continuous, and that by classifying it in phases we have inevitably simplified the picture. Just because a particular intervention can be assigned to a time when, say, walls were being painted in the Third Style, we should not think of it as a unique, compartmentalized phenomenon; the insula was constantly evolving, and it would have continued to evolve had the eruption of AD 79 not put a stop to the process. Our phases are merely a convenient way of trying to impose some kind of chronological pattern on the chain of events. In all probability changes took place more frequently and in a more piecemeal fashion than is normally realized. It is, of course, difficult to be precise about this, because our methods of dating are too unsophisticated for the necessary fine tuning; but we may suspect that, for every major rebuilding programme, there were many minor adjustments and modifications which have left no trace, or virtually no trace, in the archaeological record. And such adjustments may have happened within relatively brief periods. One has only to look at houses in modern Britain to appreciate the nature and the frequency of the alterations which they tend to undergo, ranging from the blocking or opening of windows to loft conversions and the addition of new wings. The house in which the present writer lives has been radically enlarged on two occasions, as well as having a doorway blocked, most of the windows replaced, and part of the roof redesigned—all within the space of thirty five years, and mostly in separate operations, in other words in what can technically be regarded as distinct phases. However much the archaeologist would like to divide a building’s history into broad and clear-cut chronological categories, the reality of the situation will probably more often than not have been far more complex.