What the data says

Five findings from
120,000 verses of sacred writing

These results emerged from two independent computational methods applied to 30 sacred traditions spanning 5,000 years. Neither method was designed to find these specific patterns.


01
Confirmed by two independent methods

Philokalia ↔ Zhuangzi

Greek Christian desert father asceticism and Taoist philosophical writing share no historical contact, no common vocabulary, and no common cosmology. The Philokalia speaks of hesychia — the stillness that opens the soul to divine light. Zhuangzi speaks of ziran — the spontaneous naturalness of things that do not strive. A neural network trained on human language, and a 13-axis transformation framework applied independently, both place these two traditions in the same geometric neighborhood. Neither method knew about the other's conclusion.

emb=0.843 axs=0.809Explore Philokalia →
02
The closest cross-cultural pair in the corpus

Book of the Dead ↔ Popol Vuh

Of all 435 possible tradition pairs, Egyptian funerary cosmology and Mayan creation mythology show the highest agreement between both similarity lenses. Two civilizations separated by an ocean, with zero documented contact, converge in both what they say and how they transform the reader. The journey through the Duat and the journey of Hunahpu and Xbalanque through Xibalba are, computationally, the same kind of text.

emb=0.964 axs=0.940Compare traditions →
03
Same meaning, different mechanics

Bhagavad Gita ↔ Tao Te Ching

These two texts are nearly identical in semantic space — both speak of the eternal, the natural, the non-striving. Their embedding similarity is 0.977, among the highest in the corpus. But their transformation axis profiles diverge significantly (0.623). The Gita pushes hard on action, surrender, and the individual's relationship to the divine. The Tao moves differently. Same destination in meaning space, different roads. This is the kind of distinction that escapes both conventional scholarship and surface reading.

emb=0.977 axs=0.623 Δ=+0.354Search by axis →
04
The highest similarity in the entire corpus

Guru Nanak and Rumi: 0.991

Of all 120,966 verses across 30 traditions, the devotional poetry of early Sikhism and the Persian Sufi poetry of the Masnavi are most semantically similar — similarity score 0.991. This is historically explicable: Guru Nanak was deeply shaped by Sufi thought and lived in a milieu where the two traditions were in active dialogue. But having it confirmed with a precise numerical value — and seeing it sit at the very top of the corpus-wide ranking — is a different kind of knowledge than historical narrative alone provides.

emb=0.991 axs=0.906View tradition pairs →
05
The delta matrix as proof

The gradient is real

Across all 435 tradition pairs, embedding similarity consistently exceeds axis similarity. The entire delta matrix is positive. The sacred vocabulary is universal — every tradition is talking about the same things: suffering, surrender, union, clarity, grace, the eternal. But each tradition does different transformational work with that shared vocabulary. The gradient operates beneath the surface of language. This is a precise, falsifiable, computationally grounded version of the perennial philosophy argument that has been made informally for centuries.

All 435 pairs: Δ > 0View the heatmap →
These findings are drawn from a corpus of 120,966 verses across 30 sacred traditions, analyzed using semantic embeddings and a 13-axis transformation framework.
Explore the corpus →View resonances →