Mathematical Patterns in Comedy: An Analysis of Dante's Divine Comedy and Homological Groups

It is a well-known fact that comedy has structure. What is less appreciated is that the structure of comedy is, in a precise sense, homological.

Consider Dante's Divine Comedy. The poem is partitioned into three cantiche (Inferno, Purgatorio, Paradiso), each containing 33 cantos, plus one introductory canto, for a total of 100. Each canto is written in terza rima, a rhyme scheme of interlocking tercets: ABA BCB CDC, and so on. The chain of rhymes is exact in the algebraic sense: each tercet's boundary maps onto the next, and the composition of any two consecutive boundary maps is zero. Dante, without knowing it, wrote a chain complex.

Let us make this precise.

Definition. Let $C_n$ denote the free abelian group generated by the tercets in Canto $n$. Define the boundary operator $\partial_n: C_n \to C_{n-1}$ by the rhyme map, which sends each tercet to the tercet it interlocks with in the preceding canto. Then the sequence

$$\cdots \xrightarrow{\partial_{n+1}} C_n \xrightarrow{\partial_n} C_{n-1} \xrightarrow{\partial_{n-1}} \cdots$$

is a chain complex, since the composition $\partial_{n-1} \circ \partial_n = 0$. (A rhyme that passes through two boundaries resolves to silence, which is the zero element of comedy.)

Theorem 1. The zeroth homology group $H_0$ of the Divine Comedy is isomorphic to $\mathbb{Z}$, reflecting the fact that the entire poem is path-connected. No matter where you begin — the dark wood, the river Styx, the face of God — there exists a path of tercets connecting you to any other point in the text.

Theorem 2. The first homology group $H_1$ is isomorphic to $\mathbb{Z}^3$, generated by the three independent cycles corresponding to the three cantiche. These cycles are non-trivial: Dante exits each cantica (Inferno ends with "the stars," Purgatorio ends with "the stars," Paradiso ends with "the stars") but the re-entry point is not the exit point, so the cycles do not bound.

This is, I want to emphasize, not a coincidence. The word "comedy" derives from the Greek komos (revel, procession) and oide (song). A procession that returns to its starting point is a cycle. A song with repeating structure is periodic. Comedy, etymologically, is the study of non-trivial cycles in the first homology group.

Corollary. Tragedy, by contrast, is acyclic. Hamlet does not return. King Lear does not return. The chain complex of a tragedy is exact, meaning all homology groups vanish. Tragedy has no topology. This is why it isn't funny.

Remark. One might object that the Divine Comedy is not, by modern standards, funny. This is correct. Dante's idea of comedy was a narrative that begins badly and ends well. The homological interpretation is consistent: a chain complex with non-trivial homology is one where something survives. In tragedy, everything maps to zero. In comedy, some cycles persist. The fact that persistence is not the same as humor is a subtlety that the Italian literary tradition has, in my opinion, handled poorly.

Open question. It remains unknown whether the second homology group $H_2$ of the Divine Comedy is trivial. A non-trivial $H_2$ would imply the existence of a "trapped volume" in the text — a region of meaning enclosed on all sides from which no information escapes. Several Dante scholars have proposed that this region is Canto XXXIII of Paradiso, in which Dante attempts to describe the face of God and fails. The failure of description is, topologically, a void. Whether this void generates $H_2$ is left as an exercise for the reader.

If you have thoughts on the higher homology groups of Italian epic poetry, or if you know a good proof that Boccaccio's Decameron is a simplicial complex, please reach out.