This work studies orientations of 150 parish churches with Marian dedications in the province of Soria, Spain, grouped into subsamples according to Romanesque chronology and dedication to the Marian advocation of the Assumption. The hypothesis of observing local sunrise for aligning churches versus solem orientem during a restricted Romanesque working calendar (i.e. excluding winter and harvest-time) is assessed via a large number of simulated probability densities of declinations, allowing for the application of the confidence envelope method. It was found that the hypothesis cannot be discarded for Romanesque churches with Marian dedications other than to the Assumption, alignments to sunrise on specific Marian feast days being absent in these churches. In contrast, the null hypothesis of the working calendar can be discarded for the Romanesque churches dedicated to the Virgin of the Assumption. As shown in previous studies, about 50% of these churches show intended alignments, of which two thirds are toward the sol aequinoctialis and one third to sunrise on the feast day (15 August). This last alignment pattern is a singular feature of the advocation of the Assumption, and its use in Spain cannot be older than the mid-to-late twelfth century. The theological or theoretical foundations for this practice are unknown. The paper shows that the methods used by archaeoastronomy for the quantitative statistical assessment of null hypotheses are powerful tools for inferring alignment patterns of cultural relevance.
Read full abstract