Christian churches generally point eastwards. Liturgical traditions dating from medieval times associate the direction east—the rising place of the heavenly bodies and particularly of the sun—with the resurrection of Christ and the dawning of the “day of eternity” for righteous souls. Worshipers therefore face, symbolically, their eventual home in paradise.
“East,” however, evidently did not necessarily mean “due east.” The orientations of many churches actually deviate from true east by considerable amounts. In modern times church builders were often constrained by the space available within densely populated towns and cities, so this is scarcely surprising, but in earlier centuries and in rural settings such limitations seldom existed. Given the care and attention that were afforded to many other aspects of a church’s construction, it seems inconceivable that their orientations were merely poor attempts at facing due east. The question, then, is how the direction in which a given church should face was determined in practice. What motives and procedures led to the resulting orientations is seldom clear from surviving historical accounts or liturgical texts, and the question has fascinated a number of scholars in recent years.
There have been various suggestions, the main ones of which nearly all relate in some way to the rising sun. One is that churches simply needed to face within the eastern quarter of the horizon; the details were not important. A second is that they needed to face within the solar arc—toward a point on the horizon where the sun would rise at some time in the year—but again, the actual position within this range was not important. A third suggestion is that they faced the rising sun on the day when the foundations were set out. A fourth: they faced sunrise on the feast day of the saint to whom the church was dedicated. A fifth: they faced sunrise on the equinox. And a sixth possibility: they faced sunrise at Easter. The list is not exhaustive, and it is not at all improbable that different traditions and practices held sway in different places and at different times.
Occasionally, we have historical evidence of a particular practice. For instance, there is clear support for the third suggestion in seventeenth and eighteenth century England. Historian Sir Henry Chauncy, in a topic describing the historical antiquities of Hertfordshire written at the end of the seventeenth century, reported a common practice of aligning the foundations in the direction of sunrise. Over a century later, the poet William Wordsworth graphically described a similar practice far away in the county of Westmoreland. Those who were charged with building Rydal Chapel kept vigil through a whole night, waiting for the sun to appear, at which time they carefully laid out the foundation stones and thus ensured that the building— and especially the high altar—would be correctly placed. The night concerned was not chosen arbitrarily: it was the night preceding the feast day of the church’s patron saint.
Nevertheless, it is possible that Chauncy, Wordsworth, and others were merely expressing a romanticized notion rather than hard evidence. For that, we need to measure the orientations of particular churches and compare them with sunrise on the day when construction began, or with sunrise on the patronal feast day. Sadly, the date when construction started was rarely recorded; what mattered was the date of consecration of the completed church. Identifying the patronal feast day can also give us problems, since we do not always know the saint to whom the church was originally dedicated. Even where the church is still in use, the modern patron may not be the same as the original one.
In most cases, the only evidence remaining is from the orientations of the churches themselves. Here we must be careful, since so many churches have been partially rebuilt, extended, or altered since their original construction. It is certainly necessary to pay strict attention to historical records of construction phases where they are available. How can the orientations themselves help us choose between the six possibilities already mentioned? The answer is that if we observe a large enough group of churches conforming to a common practice, then the resulting spreads of orientations should be distinctively different. At all but the most northerly latitudes (above about 55°N), general orientation within the solar arc would result in a narrower range of orientations than general orientation within the eastern quarter, but both practices would simply result in a fairly even spread of orientations within this range, perhaps concentrated toward the middle. Orientation on sunrise on the day of laying the foundations would also result in a spread of orientations throughout the solar arc, but concentrated toward the ends if the foundation dates were spread evenly through the year. This is because the change in the sun’s rising position from day to day is much smaller close to the solstices.
In order to investigate the remaining possibilities, we must determine more precisely the dates on which the sun rose at a particular spot on the horizon. For this, it is necessary to take into account the altitude of the horizon and to convert azimuths (orientations) into declinations. Orientation of sunrise on saints’ days would result in sharp spikes corresponding to sunrise on a few specific dates in the year. (Here we encounter another complication: the dates in many cases would have been determined according to the Julian calendar rather than the modern Gregorian calendar.) Orientation on equinoctial sunrise should result in a single, reasonably sharp concentration of declinations, but quite possibly significantly offset from the true (astronomical) equinox, both because of the use of the Julian calendar and depending upon how the equinox was determined in practice. Easter sunrise orientation, finally, would result in a wider spread of orientations, yet much narrower than the whole solar range, since it would correspond to sunrise over the four- or five-week period following the vernal equinox (again, reckoned in the Julian calendar for earlier churches) where Easter could fall.
In recent years, some studies of the orientation of churches of various ages in eastern Europe, Austria, Italy, and Great Britain have supported the idea of orientation upon sunrise on patron saints’ feast days. There is also evidence of solstitial and equinoctial orientation among thirteenth- and fourteenth-century churches in Hungary and elsewhere. The significance of these dates in medieval times is unquestionable: as early as the fourth century, feasts marking the births of Christ and John the Baptist had been established at the two solstices, together with feasts marking their respective conceptions (the Feast of the Annunciation and the Feast of the Visitation) at the two equinoxes. Other surveys, however, have concluded that patronal orientations never even existed. It seems that problems of data selection, together with a lack of rigorous statistics, may lie at the heart of these inconsistencies.
A recent systematic study of the orientations of medieval village churches in central England, by the American archaeoastronomer Stephen McCluskey, has revealed a complex picture that does not fit cleanly any of the six models. McCluskey restricted his sample to around a hundred churches that were dedicated to four particular patrons: Mary, John the Baptist, All Saints, and Andrew. The related feast days would be the spring equinox, the summer solstice, November 1, and November 30. While the orientations of the churches all fit comfortably within the solar arc—facing the direction of sunrise at some time in the year—they did not correlate with the four relevant saints’ days. Instead, the spread of orientations corresponded broadly to the date range from mid-February to mid-April (or in the late summer and autumn, the period from early August to early October) in the Julian calendar. This spread is too narrow to fit the “any time of year” model, yet too wide to fit the “Easter” model. There is no obvious, simple explanation, and it seems likely that a number of different traditions and practices were operating, even within this confined group of churches. However, certain preferences do stand out. For example, All Saints churches were preferentially oriented upon sunrise on the (Julian) spring equinox, something that may be explained in liturgical texts connecting the day of eternity—”god’s eighth day”—with (strictly?) eastward orientation, and with all the saints. Another, unexpected, trend is that the naves of churches of St. John the Baptist tend to point westward toward sunset on the feast of the conception of this saint on the autumnal equinox. Was this their defining characteristic?
Despite the patchiness and evident complexity of the data, there seems little doubt that astronomical observations, and particularly observations of the rising sun, were commonly used to determine the orientations of medieval churches all over Europe. Such practices may have continued well into early modern times, especially in rural areas. The investigations of such issues to date have only touched the tip of the iceberg: the evidence is extensive and remains largely unexplored.
