Is it possible to date an archaeological site by astronomical means? In theory, the answer is yes, for if we can identify an alignment and the intended astronomical target, measure where the alignment points, and then use modern astronomy to calculate where that event occurred at different times in the past, then we can fit the best date to the alignment. Stellar alignments would seem to be most promising, because the rising and setting positions of most stars change significantly over the centuries owing to precession. The rising and setting position of the sun at the solstices and the moon at the standstill limits also alter with time, although by much less, owing to the slow change in the obliquity of the ecliptic.
In practice, however, astronomical dating is rife with pitfalls. The main one is that we rarely have reliable (e.g., historical) evidence to tell us either that a particular alignment was actually deliberate, or what is was aligned upon in the first place. Where we do have historical evidence, we are likely to know the date fairly accurately already. (A case where we do have some historical evidence, but not as much as we might like, and the alignment does tell us something, is the Venus alignment of the Governor’s Palace at the Maya city of Uxmal.)
More often, we are dealing with prehistoric structures where we have no evidence other than the alignments themselves. If we find an alignment that we suspect to be stellar, we can try different stars and different dates to see if any combination fits particularly well. The problem here is that if, say, we are willing to consider the fifteen brightest stars and a five-hundred-year date range, then there is approximately a one-in-three chance that we will be able to find a star and a date to fit any alignment. It is frighteningly easy, then, to fit a star and a date fortuitously and made all the easier when we consider that only rarely does one single alignment at a particular monument stand out as the obvious astronomical candidate. In order to be fair with the data—one of the most fundamental methodological principles—we should consider all possible alignments. Added to this is the problem of extinction (the dimming of a star at low altitude due to the earth’s atmosphere), which may mean that most if not all of our fifteen stars wouldn’t have been visible all the way down to the horizon in the first place. If we are willing to postulate (as some people have done) that structures were aligned upon the point of appearance or disappearance of a star rather than its actual rising and setting point, it increases the chances of our being able to fit a date and star fortuitously to any given alignment (especially at high latitudes, where the astronomical bodies rise and set at a fairly shallow angle). Overall, the potential for circular argument is obvious.
Yet despite these problems, there have been cases where postulated stellar alignments can help in the process of dating a site. One relates to the Pyramids of Giza in Egypt and another to a perplexing later prehistoric sanctuary in Mallorca, Son Mas.
Solar and lunar alignments don’t suffer from these problems but have others of their own, mainly that the change over time is very small. The change in the declination of the solstitial sun, for example, amounts to only about one arc minute per century. This means that only if we have good reason to suspect that a solstitial alignment had very high precision, such as that envisaged by Alexander Thom at British megalithic sites, can such an alignment be used to indicate a date accurate to within a few centuries. Even then, variations in refraction may provide an insurmountable obstacle.
