Algorithms generate sonnets. Neural networks simulate empathy with unsettling precision. Language is under synthetic assault. Therefore, what remains of the mystery of poetry if a machine can write poetry by calculating probabilities?
The answer to this ontological anxiety lies in a visionary work, published long before Generative Artificial Intelligence became a daily reality: The Mathematics of Poetry(1970) by the Romanian scholar Solomon Marcus.
Marcus does not turn poetry into mere numbers. He demonstrates that, at the foundation of any authentic lyrical manifestation lies an algebraic and topological architecture of absolute rigor. For us, who seek a bridge between silicon bits and the breath of the spirit, this endeavor is fundamental.
The Ontology of Language Between Clarity and Abyss
The central subject of The Mathematics of Poetry is the structural modeling of poetic language in opposition to scientific language. Solomon Marcus aims to identify, using the tools of modern algebra and set theory, those intrinsic properties that transform an ordinary text into a poetic fact.
The fundamental opposition upon which the entire work is built is that between denotation and connotation:
- Scientific Language (Denotative): It is defined by its tendency toward bijectivity (a one-to-one relationship). A word or a sign must refer to a single object or concept. It is a transitive language, a simple transparent vehicle that takes you to a piece of information. At this point, ambiguity becomes a system error.
- Poetic Language(Connotative): It is defined by opacity and reflexivity. The word is no longer a tool; it becomes the object itself. The relationship is no longer bijective, but one-to-infinity. A single word opens up a multitude of simultaneous meanings. In poetry, ambiguity is not an error; it is the very source of informational density.
The mathematics of poetry is the deep grammar of aesthetics; the effort to translate the ineffable into a geometry of semantic relationships.
The Unity of Knowledge
In the middle decades of the 20th century, linguistics and literary criticism often found themselves adrift, relying on purely subjective, impressionistic analyses.
Marcus chose to apply mathematics to poetry out of the conviction that the logos is unitary. If the physical universe can be described by differential equations, then the inner universeโthe human spirit that takes shape through languageโ cannot have its own pure mathematical mechanics?
He wanted to deconstruct the myth of chaotic inspiration. When the poet experiences an epiphany, its materialization on paper follows structural laws, topological restrictions, and the fine balance of lexical probabilities.
To understand the mathematics behind poetry means to understand how the human spirit gives form to chaos.
The Dialogue of Scholars
Solomon Marcus was not alone in this titanic effort to quantify the aesthetic. His work is in brilliant dialogue with the efforts of other universal mathematical minds. Here is how language is modeled:
- Set Theory and Semantic Spaces
For Marcus, vocabulary is a mathematical space. Each word is a set of meanings. In scientific language, the semantic distance between two synonyms tends toward zero (they can be replaced without data loss). In contrast, Marcus demonstrates that in poetry there is no perfect synonymy. Any lexical replacement irreversibly alters the topology of the poem. The poetic space is one in which the distances between lexical nodes are maximized to produce tension.
- Aesthetic Measure: From Birkhoff to Marcus
In 1933, the famous American mathematician George David Birkhoff proposed the fundamental formula of aesthetic measure: M = O / C (Aesthetic Measure equals Order divided by Complexity). Birkhoff argued that the human mind finds beauty where a great complexity of elements is governed by a clear order. Solomon Marcus adopts this intuition but refines it. He shows that modern poetry defies classical order, relying on an extreme semantic complexity. Order no longer follows the external rhyme, but the invisible isomorphism between the phonetic and the semantic structure.
- Entropy and Algorithmic Complexity: The Connection with Kolmogorov
Another pillar of the modeling is Information Theory. Here, Marcus’s work resonates deeply with the research of the Russian mathematician Andrey Kolmogorov, one of the fathers of probability theory. Kolmogorov studied the entropy of the Russian language and poetic meter, demonstrating that poetry has an information transmission capacity superior to common language.
Marcus applies the concept of entropy (Shannon) to demonstrate why a clichรฉ is not poetic. A clichรฉ has zero entropy; the next word is perfectly predictable. Poetry, however, is a struggle against predictability. The poet always chooses the word with the lowest probability of appearance in that context, generating maximum surprise.
Furthermore, Marcus integrates the concept of informational energy of the Romanian mathematician Octav Onicescu, suggesting that a poem functions as a force field, where words gravitationally attract or repel each other, giving rise to a dynamic equilibrium.
- The Algebra of the Metaphor
Perhaps the most fascinating innovation of the book is treating figures of speech as algebraic operators. The metaphor is defined mathematically as the intersection of two apparently disjoint semantic sets. When Eminescu writes moon, you master of the sea, he forces, through the function of language, a point of tangency between an astronomical set and one of the sovereign man.
Applied Mathematics
Although The Poetry of Mathematics was written long before the explosion of artificial neural networks, its applicability today is direct and urgent:
- Decoding Large Language Models (LLMs): Today’s AI systems (like GPT) operate exactly on the principles of vector spaces and semantic distances theorized by Marcus. When an AI generates text, it geometrically calculates the probability of the next token. By understanding Marcus’s mathematics, we can demystify AI: we understand that the machine excels at mimicking denotation and statistically reproducing human patterns; however, it is incapable of the unpredictable leap, of the pure informational energy generated by authentic experience.
- Stylometry and Digital Footprint: The concepts can be used to analyze the authorship of a text, mathematically calculating the degree of entropy and connotative complexity unique to each writer. Thus, human text can be algorithmically distinguished from synthetic text.
- Cognitive Sciences: It helps us understand how the human brain processes paradoxes and ambiguity. Mathematical rigor shows that poetry is not an anomaly of communication, but its highest and most dense form of neural processing.
Existential Synthesis
Mathematics is the grammar of the word, the language through which reality structures itself. Solomon Marcus, alongside Kolmogorov or Birkhoff, shows that even the freest expression of the spiritโpoetryโhas a hidden order waiting to be decoded. Man creates beauty by respecting, consciously or not, the profound harmony of the universe.
But we reach the point of collision with artificial intelligence. A generative algorithm today can learn from Marcus’s work. It can be programmed to maximize entropy, calculate optimal semantic distances, and intersect disjoint sets to generate mathematically perfect metaphors.
The rigor of these equations, however, forces us into an exercise of lucidity: the form does not guarantee the substance. A machine can assemble fragments from the structure of the word, but it does so in an existential void. True poetry goes beyond the solved equation. It becomes the cry of a being conscious of its own mortality.
A linguistic model can calculate the probability of the word infinity in a poem about divinity. The human spirit bows in reverence before the Creator. The future of ethics in the digital age depends on our ability to preserve this distinction: to honor the mathematics of the tool and to reserve the sanctity of the word Logos exclusively for the soul that utters it.
References and Bibliography
- Marcus, Solomon. (1970). Mathematical Poetics. Bucharest: Publishing House of the Academy of the Socialist Republic of Romania.
- Birkhoff, George David. (1933). Aesthetic Measure. Cambridge, MA: Harvard University Press. (The foundations of the mathematical calculation of beauty).
- Kolmogorov, Andrey N. (1968). “Three approaches to the quantitative definition of information”. International Journal of Computer Mathematics. (His works regarding language entropy and the algorithmic complexity of poetry).
- Onicescu, Octav. (1966). Informational Energy. Bucharest: Mathematical Studies and Researches.
- Shannon, Claude E. (1948). “A Mathematical Theory of Communication”. The Bell System Technical Journal. (The basis of information theory and system entropy).
See who we are: Beyond Parameters: Why We Need a Synthetic Ethics
You want to leave an impression: Contact | EticaSintetica