In medieval Europe, Gregorian chant was everywhere. Its unaccompanied melodies were sung daily in huge cathedrals, parish churches, remote monasteries; it was an omnipresent musical expression of medieval religious life. It was also “globalized” across Europe, after Charlemagne demanded standardization of worship in his new empire after 800 AD. As a sacred tradition, it was strictly conserved. Yet, the melodies that came down to us from medieval manuscripts are rarely exact copies. Throughout the centuries, chant was surreptitiously changing, and by the mid-16th century, the Church decided that Gregorian chant should be strictly standardized — again.

Our project experiments with applying phylogenetic methods from bioinformatics in order to characterize this process of diversification. We aim to determine whether the variability seen in Gregorian melodies is systematic, or result merely of transmission noise (with memorization and manuscript copying both susceptible to mistakes). We show that the bioinformatics approach is indeed applicable, directly on the raw melodies. On our pilot Christmas data, the inferred phylogeny conforms very closely to musicological expectation, it shows that Gregorian melodic dialects indeed do exist, and in an example of what computational scholarship can bring to the humanities, led to re-evaluating existing knowledge about one of the sources.