Theoretical Structural Archaeology
by Geoff Carter
Stonehenge was a building. That’s it, no mystery.
If it was a rectangle this would not be an issue, but British Prehistoric buildings are predominantly circular from this period onward.
The rings of postholes at Stonehenge [Y, Z, Q, and R holes] are often ignored, or are thought to be redundant stone holes, but it is just one of a group of concentric timber structures known from various periods in British Prehistory. Like Woodhenge, Durrington Walls, Mount Pleasant, and The Sanctuary, Stonhenge was a large timber building. This was tentatively recognised by Tim Darvil in 1996, who called them Class Ei structures.
What I am arguing:
A: Stonehenge is a class Ei Timber structure.
B: The archaeological footprints of class E structures share geometric, proportional and structural features, all of which are consistent with their being roofed buildings.
While each of these structures is unique, they are sufficiently similar in their spatial arrangements, in over a dozen ways, that are best explained by the requirements of architectural solutions directed at supporting a timber roof. All these structures have been discussed in earlier articles.
These arguments are made on the basis of the spatial distribution and relationships of the archaeological plans, and are demonstrated by accurate technical diagrams.
The twelve reasons why Stonehenge was a building
1. Context: For north-western Europe, a wet windy temperate zone with cold winters, for viable mixed sedentary agriculture, a robust and complex built environment is required, as well as a range of specialised skills and technologies to manufacture and maintain it. From the Neolithic period onwards, and all through prehistory into historical periods, posts have been set in the ground as the foundation of buildings and other aspects of the built environment, such as fortifications. This is how configurations of postholes of similar size and depth are normally understood and interpreted in European archaeology.
2. Form: These Class Ei structures are based on a system of 5 rows of posts; this is the form of roof evident in the Neolithic. All of the observations below are consistent with an attempt to create a roof based on this model. Each of these round buildings is built like a Neolithic long house turned in a circle. They are annular, with the tighter circles probably having been covered with a central cone. Not all have 5 rings of posts: Stonehenge has 4 plus a circular wall, while Woodhenge and Durrington Walls have additional rings on the inside of the circle.
3. Later examples: These Class Ei structures are known from the late Iron Age, as at Naven Fort, dated to where it has a proto historical context. The pattern is also evident in the layout of the Dark Age ringfort at Lissue. This clearly argues that concentric rings of posts are a technological solution to roofing circular spaces. Stonehenge is simply a particularly unique variation on a type of large building.
4. Proportion: For a roof of this form to ‘work’, it must be proportional and be composed of parallel elements at different heights. Viewed in section, this is broadly the case:
Line of posts central to an inner and outer ring.
Flanked posts for an arcade [not at Stonehenge, however]
Each rafter pair is nominally supported by five horizontal parallel timbers [10 posts], with a tie running across the base.
The complexities of this geometry are covered by reasons below [Interlace theory].
5. Posthole depth: In this form of roof, the ridge at the top is supported by posts, and this explains why the central row of posts is usually the largest and deepest. This does not preclude a set of larger posts on the inside of this, depending on how the roof in the centre of the structure was fashioned.
6. Scale: Class Ei buildings are about 50’/15m across the roof; this is the traditional width for large timber buildings in Britain. Much subsequent development in architectural technology was directed towards pushing these limits, and creating space less encumbered with posts, perhaps reaching one climax with Westminster Hall, [68’(20.72m)].
7. Oak timber building: The growth pattern of young oak trees dictates the availability of suitable timber to form continuous timber components like rafters and ties, when the tree is neither too thick at one end, nor too thin at the other. For practical purposes we can think of a 50’ /15m tree trunk tapering, this being the best timber commonly available. 
Some individual timbers may be considerably longer, up to and even over 60’; these can be split from the trunks of larger, more mature trees, just as in medieval buildings.
Prehistoric builders may have had an advantage in terms of natural forest resources.
This limit is also reflected in the size of roundhouses.
Trees of this size/age range are represented in the diameter of postholes.
8. Posthole diameter: [Tapering ties]: Postholes in the outer ring are larger than those in the inner ring, because the lowest horizontal timber, a 50’ tie, is thick at one end and thin at the other, and this is reflected in the size the post [hole] that supports it. These structures always have their smallest posts in the centers.
9. Structural detail: Spans: So it can be supported, the path of each of the horizontal ties across the base of the roof, directly between the corresponding inner and outer posts, will pass next to, but not through, the intervening post. The maximum unsupported timber does not exceed c.25’. At Stonehenge the ties would be supported by the center posts of the Z holes, and then by the Sarsen ring.
10. The Sarsen Ring: This is a perfectly serviceable load-bearing wall. It is:
Engineered to a high standard
Positioned like an arcade
Implies the presence of a load [roof]
11. The Trilithons: These are a load-bearing post and lintel component;
Positioned at the centre of gravity of a conical roof based on the inner post ring [R].
12. Woodhenge: This is not a circular building, but it is the exception that proves the rule: it is based on a Lozenge, a difficult and singular shape to roof, yet it still works as a building. This geometry is shared with the Bush Barrow Lozenge and other motifs of this period.
Once it is understood that these structures are buildings, further consideration of their scale and circular geometry makes it apparent that they cannot be simplistic in design. Accurate three-dimensional modelling and testing has to consider more complex issues, and Interlace Theory was developed to model and understand structures with curving roofs supported by post rings.
13. Structural geometry: [Interlace Theory]; Interlace properties: By joining the posts of a ring to each other at different intervals, the parallel elements at different heights and angles necessary to facilitate a curving multilevel roof are created. Longer elements are higher on the widening outside surface of the roof, but lower on the inner surface. It also increases the rigidity and strength of the structure by:
Creating a series of interlocking polygons
Interlock with the next post ring
14.Assembly: At its most basic, each section of roof can be thought of as five parallel pieces of wood supporting a rafter pair, each element supported by posts in the post rings. They could be assembled by starting from the lowest section of roof; from a pair of posts in the outer ring and a set of parallel elements in the inner ring. The adjacent sections sharing one of the original posts would have to be slightly higher, and so on until a high side is reached.
As you have to start with a tie, the intervening posts that support it would have to be in place first.
The inner ring[s] and roof is a scaled down version [reflection] of the outer roof shape.
It is possible to build such a structure in two directions starting at a low point.
It ensures that each part of the roof is at its own unique height preventing special conflicts in a roof that turns in on itself.
This implies that these structures had a higher side, and therefore the high part of one post ring probably roughly corresponded to the lower posts of the next ring, and so on, allowing for a continuous build in the horizontal timbers.
Think about bricklaying or basketry: This is a system of assembly for tree trunks, joining the posts together continuously, like '70s string art, using timber instead of string, and posts rather than nails.
Stonehenge is unique, but so are all the other structures discussed. However, despite this, they have a series of features in common which could not be coincidental, and that are entirely consistent with the proposition that class Ei buildings were roofed. There are still more reasons I could give, but it would become abstracted and technical. The Advanced Considerations are already well into architectural model making.
There is no archaeological evidence that these structures were not buildings, only a prejudice in existing narrative. However, a lack of structural understanding and appropriate research to resolve the issue cannot serve as argument that these particular assemblages of postholes are the product of a unique, unprecedented, and undocumented religious ritual created to explain ‘Timber Circles’. Nor is circularity sufficient ground to infer a simplistic relationship with the Stone Circles of earlier periods.
Clearly, this challenges the current academic narrative, with its emphasis on the perceptions, belief, rituals, and cosmologies of ancient peoples, none of which is implicit in the evidence that is recovered by archaeology.
Reverse engineering structures works by deduction: building models that work, discarding the ones that don’t. This research did not set out to prove anything, and certainly did not set out to understand Stonehenge and other Ei buildings. That has been a by-product of testing ideas about postholes that started as an enquiry into Iron Age and Roman timber buildings back in 1990.
The scale of these structures clearly implies they were built at the limit of what was considered prudent by builders in Prehistory. This is a craft that was several thousand years old, and by this period we have evidence of a new elite who would normally be expected to express wealth and power in the built environment.
My presumption is that Stonehenge was a temple built house; the Bluestones, the unprecedented use of a stone load-bearing wall; and pillars in the centre a technological approach reminiscent of Mediterranean Europe, suggesting imported craftsmen. But this is only the very pointy end of a much larger wedge.
The real interest lies in the many other buildings that can be understood in detail, and the new light this can shed into the nature of the prehistoric built environment.
The detailed understanding of how these structures worked and were assembled, briefly summarized above, allows the building of a virtual timber-by-timber reconstruction using CAD technology, and I will be discussing this in my paper at the CAA 2012 Conference at Southampton.