Some of the ideas discussed in this blog are published in my book called "The Bluestone Enigma" -- available by post and through good bookshops everywhere. Bad bookshops might not have it....
To order, click

Wednesday, 24 August 2016

Bang goes the Carngoedog Neolithic quarryman's row........

 Carn Goedog.  The postulated quarrying settlement is right at the base of the long rocky slope, at the right hand edge of the photo.

A couple of years ago there was considerable speculation from Prof MPP and others that the workforce of the fanciful and fantastical Carn Goedog Neolithic bluestone quarry was housed in the "row of huts" located at the foot of the crags, on the north side. It was speculated that because there were clear rectangular patterns visible on the ground the huts were not likely to have been Bronze Age, but maybe Neolithic instead. A quarry plus accommodation for the workers -- a very pleasing scenario.......

Then it went all quiet, and we now know why. I get sent information (including unpublished reports) all the time, and I really don't know where this one came from. It's a draft or unpublished report with the following title:

Carn Goedog medieval house and settlement

By Rhiannon Comeau, Rob Ixer, Mike Parker Pearson, Duncan Schlee and Kate Welham

The fieldwork might be part of an unpublished thesis by Rhiannon Comeau. She and the others examined the settlement site, and the discussion is appended below, with acknowledgement.

The result of the work? The settlement is essentially medieval, and has nothing to do with the Neolithic and nothing to do with quarrying. Some of the traces of older settlements might go back to the Bronze Age, but not earlier.

See also:



All of the artefacts from the excavation of House C derive from layers immediately outside it and are presumably the remains of refuse discarded from its interior. They point to the house being occupied in the high and/or late medieval period and subsequently abandoned. However, earlier occupation cannot be ruled out given the aceramic character of the local area before the 12th century, and the notoriously poor survival of pre-Conquest material culture. There were no artefacts that could be considered to be either earlier or later in date, and no evidence of multiple periods of occupation, suggesting that House C was built, used and abandoned within the medieval period rather than being reoccupied at different periods in the past. A radiocarbon date from the floor deposit could establish the date of occupation of the building. 

Similarities in the form of the neighbouring buildings suggest that Houses A-B and D-J are likely also to date to the medieval period. The group perhaps represents a havod (a small seasonally occupied dwelling associated with exploitation of the upland pastures). The curvilinear houses I, K, L, M, N (and perhaps O) are undated but the more denuded appearance of their stone walls may well reflect their greater antiquity as pre-medieval structures. Remains of field walls forming a north-south coaxial pattern leading northwards onto the lower ground north of Carn Goedog could be associated with either or both of these settlement clusters.

In summary, then, we have here a group of nine sub-rectangular buildings and ancillary structures. Their lengths, when surveyed before excavation, ranged from 4.4m to 7m externally, with most having unexcavated internal dimensions of around 4m x 2-3m. All have their long axis at right angles to the contours. House C, whose pre-excavation survey measurements were 5m x 4.6m externally and 3.6m x 3m internally, was revealed by excavation to have external measurements (at maximum) of 6.2m x 4.4m and internal measurements of 4.5m x 2.2m. Dating evidence – pottery – suggests occupation in the medieval period, but not later.

Similar groups of single-celled habitations (described both as ‘platform houses’ and as ‘long huts’ – definitions vary) are recorded elsewhere in many upland locations in Wales (Roberts 2006: 172–5, 207). Entrances are usually on the long sides and are often – as in House C – opposed, which may be linked with occasional suggestions of internal partitions (Leighton 2012: 120; Locock 2006: 45; Silvester 2006: 34). They are commonly associated with medieval and early post-medieval occupation, although few have been excavated and even fewer are dated (Leighton 2012: 127). A cross-contour placement, at right angles to the slope, is typical of medieval examples (Roberts 2006: 177). Sizes vary considerably: the Carn Goedog long huts fall into the 4m–7m length range noted for many long huts in central Wales, but much larger examples exist, like those at Gelligaer in Glamorgan which are dated to the late 13th or early 14th century (Fox 1939: 173; Silvester 2006: 34). They are generally thought to be linked to the use of seasonal pasturage, although interpretation of evidence for the seasonal or permanent occupation of excavated dwellings is contested (Silvester 2006: 33-4). It is these seasonal pastures that are referred to as the ‘hafod’ in medieval sources; the term transfers to dwellings from the 16th century onwards (Davies 1980: 3–7; Sambrook 2006: 95–9). Thirteenth-century Welsh law uses the term ‘haf ty’ (‘summer house’) for the dwellings on the hafod, and their relatively insubstantial nature can be seen in the low compensation values attached to them (Davies 1980: 7; Jenkins 1990: 190, 353). The same law code indicates that it was expected practice for the bondsmen and animals of farming settlements to relocate to seasonal pastures from the beginning of May until the harvest was in (Jenkins 1990: 40, 236).

The location itself is also characteristic of ‘hafod’ sites, and is typified elsewhere in south Wales as ‘along a track running diagonally up a slope, in a sheltered position below the summit area’ (Locock 2006: 45). The track next to the Carn Geodog dwellings is shown on a map of Pembrokeshire of 1602 in which it links the farmland of Whitchurch to the north with the parishes of Mynachlogddu and Maenclochog on the southern flanks of the Preseli hills (Owen 1602).

As a building type, House C can thus be seen as typical of many undated upland structures, and its occupation evidence is therefore valuable, both for the unequivocal medieval date of its pottery and for the hints of gendered occupation provided by the spindle whorl. If this is a seasonal structure, this evidence is particularly interesting since the social composition of seasonal occupation could, in principle, range from whole families (the interpretation often surmised from medieval Welsh law) to the teenage boys mentioned locally c.1600 and the hired herdsmen known elsewhere in the post-medieval period (Miles 1994: 44–5; Ward 1997: 104–6; cf Fox 2012: 47ff)

There is, however, little in the site’s excavated evidence that indicates its annual pattern of occupation, whether seasonal or all-year round. The presence of pottery cannot in itself be taken as an indicator, since the site lies on the edge of what appears to have been a well-used medieval track and less than an hour’s walk from areas of permanent settlement. Seasonal land usage is suggested by the nature of the surrounding terrain as well as by its longstanding common land identity and seasonal pasture rights still exercised by farms around the common edge. These rights are thought to derive from a mid-13th century charter, granted to an aristocratic Welsh kin-group who owned much of the arable land to the north in the medieval period (Jones 1979: 28; Owen 1862: 48). Carn Goedog lies within the area defined by the 13th-century charter, and its habitations were presumably occupied by these aristocrats’ bondsmen or tenants. 

The medieval occupation indicated by the pottery is paralleled by documentary indications that any family-based seasonal movements must have ceased by the late 16th century, since they are not mentioned in contemporary descriptions of local agricultural practices (Miles 1994: 62–77, 175). The late 16th-century records of ‘hafod’ place-names for nearby permanently occupied farmsteads at Hafod Wynog (1598) and Hafod Tydfil (1585) – the latter an enclosed island of fields amidst the moors to the northwest of the site – chart accompanying changes in farming practices which see the enclosure of some areas of seasonal pasture for all-year farms (Charles 1992: 105). Other – undated – enclosures can be identified on the moorland to the north and northeast of the site in areas that are now waterlogged and, whilst these may be prehistoric, it is possible that these may also represent intensified exploitation in benign climatic intervals during the medieval period.


T said...

This is the way the world ends, not with a bang but a whimper.


Oh dear, Mike.

Alex Gee said...

"Oh Dear" Neatly summarises the whole quarrying hypothesis.

TonyH said...

"Oh dear" is, in fact, an oft - used MPP expression.

Dave Maynard said...

Stratigraphically, this does not demolish the theory that several millennia before, there was a working quarry. Simply at a later date, parts of the area got re-developed into a hafod / summer residence. Happens all the time.


BRIAN JOHN said...

That much is self-evident, Dave. But the diggers would have loved it if the settlement site could have been shown to be Neolithic, since they actually have no evidence of quarrying at all...... assumptions and speculations, but no hard data.