We introduce a method for treating a corpus of text as a semantic trajectory—-a path through high-dimensional embedding space that evolves in time as each new verse accumulates contextual meaning from its predecessors. Applying this method to the King James Bible (31,100 verses) and to a parallel Arabic corpus of the four scriptures of Islamic tradition—-Torah, Psalms, Gospels, and Quran (18,324 verses)—-we discover a striking structural asymmetry. In the KJV, the Psalms function as a semantic centre of gravity: by the Psalter, every thematic basin in the corpus has been visited, and the New Testament is almost entirely composed of returns (عودة) to previously traversed territory. No Old Testament—New Testament rupture appears. In the Arabic corpus, by contrast, each scripture occupies its own distinct region of embedding space with near-zero cross-scripture overlap, even when linguistic barriers are removed. We argue that this asymmetry is an artefact not of theology but of translation: the KJV's single authorial voice flattens the heterogeneous source texts into a unified register, producing a semantic coherence that is a property of the translation, not of the original. The embedding space sees the translator, not the scripture. This finding vindicates Harold Bloom's thesis that the KJV is a literary entity sui generis—-``the English Qur'an''—-a foundational text for the English language that is categorically distinct from the Hebrew and Greek originals it translates. Full HTML text: https://icra.tanazur.org/papers/semantic-topology-translation/ — ICRA pre-print series: https://icra.tanazur.org/
Poernomo et al. (Mon,) studied this question.