We are grateful for the opportunity to respond to Euan MacKie's paper above, noting also the longer version on the ANTIQUITY website. We have restricted ourselves to general points.
As regards methodology, we agree absolutely that conclusions involve balancing probabilities and that in academic debate there is a fundamental expectation that we all strive to `play fair with the evidence and not to treat it partially'. A consequence of this is a need to present others' arguments fairly when criticizing them, something we strove to achieve in the original article. On the other hand, MacKie makes the remark `Ruggles considers that his own fieldwork has disproved the existence of such alignments [i.e. long sightlines to distant natural foresights that could be used for astronomical observations of high precision], and that any confirmation claimed from independent tests of the idea is mistaken'. This is both emotive and misleading. A few points must suffice here.
1 It can never be stated categorically that long sightlines to distant natural foresights did not exist, and Ruggles has never claimed this. Absence of proof is not the same as proof of absence, as some commentators (e.g. Ashmore 1999: 27-8) have been careful to emphasize. What was shown in the early work by Ruggles was that the statistical evidence put forward by Thom, which ostensibly demonstrated a widespread and consistent practice of high-precision solar and lunar alignments all over Later Neolithic Britain, was completely undermined by selection effects in the data. Furthermore, lunar alignments of the precision claimed by Thom in his later publications are simply impracticable (see a string of publications from Heggie 1972 onwards, cited in Ruggles 1999: 61-2), and more recent work on variable atmospheric refraction implies that: the same is true of high-precision solstitial markers (Schaefer 1993; Schaefer & Liller 1990). We cannot rule out the existence of individual sightlines of remarkable precision, if not quite at the level envisaged by Thom, but the problem is to find a secure theoretical and methodological basis for distinguishing intentional, meaningful configurations of monuments and natural topographic features in any particular instance from what are often a great many possibilities.
2 Despite what MacKie has said before (MacKie 1986) and repeats here, it was not the fieldwork in western Scotland published in 1984 (Ruggles 1984a) that critiqued Thom's evidence for foresights of a precision greater than about 6 arc minutes, but three preceding papers (Ruggles 1981; 1982; 1983). It was in the light of this earlier work that the wider project, attempting to move on constructively from the earlier critiques, limited the precision to about 6 arc minutes and concentrated on `indicated horizon ranges' rather than putative foresights. This has been stated clearly in the literature on several occasions (e.g. Ruggles 1984b; 1989 and the evidence is presented step by step in chapter 2 of Ruggles 1999).
3 It is not unreasonable to look systematically for instances where a single, particularly prominent feature, such as a `clear-cut notch', stands out in a target range of horizon; and indeed, one of the main criticisms of Thom's own alleged high-precision alignments was that the great majority were not actually like this (see Ruggles 1981; 1982; 1983). However, one should be cautious of `one-offs' like the Sornach Coir Fhinn example as they may have arisen fortuitously. For a approach along these lines, see the systematic studies of axial stone circles and stone rows in southwest Ireland undertaken by Ruggles in the mid 1990s (Ruggles 1994; 1996; Ruggles & Prendergast 1996). While they may include individual instances of alignments on equinoxes or mid-quarter days, these and other studies do not reveal systematic evidence for a precise Thomian calendar but broader patterns of relationship with the sun or moon amongst local groups of monuments. …