Every song you’ve ever heard is part of one giant, blurry spectrum rather than a collection of distinct musical shapes.
April 16, 2026
Original Paper
Melodic contour does not cluster: Reconsidering contour typology
arXiv · 2604.13119
The Takeaway
We used to think music was built from a specific 'alphabet' of shapes—rising, falling, or arching melodies that we could neatly categorize. But when researchers analyzed thousands of folksongs and chants, they found these categories don't actually exist in the data. Instead of discrete types, melodies flow into each other like a continuous gradient. This means the way we teach and analyze music as a series of 'types' is more of a human-imposed grid than a biological reality. For you, it means your brain isn't just recognizing 'sad falling' or 'happy rising' shapes, but navigating an infinite, messy landscape of sound.
From the abstract
How to describe the shape of a melodic phrase? Scholars have often relied on typologies with a small set of contour types. We question their adequacy: we find no evidence that phrase contours cluster into discrete types, neither in German or Chinese folksongs, nor in Gregorian chant. The test for clustering we propose applies the dist-dip test of multimodality after a UMAP dimensionality reduction. The test correctly identifies clustering in a synthetic dataset, but not in actual phrase contours