In medieval Europe, Gregorian chant was everywhere. Its unaccompanied melodies were sung daily in huge cathedrals and small parish churches, in cities and remote monasteries; it was an omnipresent musical expression of medieval religious life, and the memorization technique for passing sacred texts to new generations. It was also “globalized” across Europe, after Charlemagne demanded standardization of worship in his new empire after 800 AD. Chant transcended political and geographical boundaries as a sacred tradition that was supposed to be strictly maintained. Yet, the melodies that came down to us from medieval manuscripts are in fact rarely exact copies. Throughout the centuries, chant was surreptitiously changing! (It had branched out so much that in the mid-16th century, the Church decided that Gregorian chant was supposed to be strictly standardized again.)
What was really going on with chant melodies in the Middle Ages? How did they evolve, even though they were not supposed to? Did some local traditions emerge? Could one perhaps reliably determine whether a chant melody sounds “more English than Bohemian”? Is it possible to tell older from newer melodies? What does this tell us about our own period of cultural globalization?
Fortunately, we have more than 13,000 chant melodies available digitally. They come from more than 600 liturgical books spanning the 11th to the 16th century, across much of Europe. Perhaps we could then answer such questions with the help of bottom-up computational methods, essentially letting the melodies “speak for themselves“? Since we think that melodies might adapt to their social and cultural environments similarly to how animals or plants evolve, we borrow our inspiration from the life sciences, and examine Gregorian chant melodies with algorithms originally developed for bioinformatics.