3.2 Post-excavation - groupings

• 3.2.1 Processual interpretative cycle of 'small phase events'
• 3.2.2 From description to' processual' interpretation

It is important to understand that excavation and post-excavation analysis work in opposite directions: while one excavates down a sequence, analysis and grouping starts always at the bottom and proceeds upwards. The end result is to structure the data into phases or periods of site events. However, for complicated and large excavations, such grouping cannot be established in one go. The more complex an excavation the more important it is to approach the task in steps. As the stratigraphic matrix of single contexts appears like looking at stars in the sky, they have to be interpreted and grouped at different levels of interpretation: at context level the interpretation is basic. Each higher grouping step broadens the interpretation: for instance, a posthole context joined with others becomes part of construction; the construction of a building; and the building was part of a regional development. Eventually all context records including their finds and samples are presented as phases or periods of development in chronological order. In an urban environment with many years of occupation resulting in a high build-up and re-working of stratigraphy, the finds within a deposit do not necessarily reflect the date of their deposit therefore they have to be regarded in sequence (Figure 2).

The convincing point of stratigraphic structuring lies in the transparency of the procedure: if anyone came to doubt the results, they can go back step by step the analytical ladder and pinpoint exactly where they think the arguments were floored.

The two main reasons for grouping therefore are:

• To establish from the basic interpretation of single context records a chain of higher levels of interpretation that links eventually all contexts chronologically.
• This interpreted and grouped site sequence provides a frame for all associated finds and samples, proving whether they were contemporary, residual or intrusive. A stratigraphically tightly defined find could also be used to shorten its known date range. However, not all occupation proceeds in clear phases, for instance, in one part of the site several phases of construction could have followed each other while in other parts an open area did not change' during several phases. Grouping also helps to make this relation obvious (Figure 3).

To make the complexity of events transparent and to avoid jumping to conclusions the conversion from description to interpretation proceeds in steps: single contexts are combined to sub-groups; sub-groups to groups and these to sub-phases/phases of land use or periods. The complexity of the grouping steps applied depends on the complexity of a site and the amount of data to be published together. Such a hierarchical way of structuring suits the organisation of the database. During grouping it is important to strictly observe only stratigraphic relationships. Each step onto a higher hierarchical grouping level is stratigraphically less secure but gains a higher interpretative plane.