The Voynich Manuscript is one of history's most enigmatic medieval artifacts, written in an unknown alphabet and a language that has historically resisted all phonetic translation attempts. Previous studies based on information theory (textual entropy) have already demonstrated that the manuscript is not a random sequence of symbols, but contains a coherent semantic signal. Building upon this established premise, the present study goes further by questioning the structural architecture of this signal: is the Voynich written in spontaneous narrative prose, or does it hide a more rigid and conditional logic? The intuition behind this research stems from the observation of a profound statistical anomaly, historically dismissed as mere copying errors: the exceptional frequency of identical strings repeated consecutively (the "double words"). By applying Machine Learning algorithms (Logistic Regression) and cross-referencing the data with Latent Semantic Analysis (LSA), this research demonstrates that these doubles are not errors, but mechanical syntactic operators. They act as structural delimiters designed to close data blocks or partition thematic sections. A 1000-iteration Permutation Test incontrovertibly confirms the mathematical causality of specific "Trigger Words" in precipitating the formation of such syntactic markers (p=0.0010). Furthermore, physical investigation of the artifact reveals that the most extreme spatial concentrations of these "logic gates" occur exactly adjacent to the 14 folios historically removed from the manuscript. This codicological anomaly supports the hypothesis of the targeted removal of conversion matrices or glossaries (Tabulae), the absence of which has rendered the surrounding complex logical transitions incomprehensible. The results of this study offer a decisive paradigm shift: the Voynich Manuscript should not be approached as a linear phonetic language, but rather as a modular data structure, whose architecture is closer to a recipe book of rigidly constrained formulas than to spontaneous human language.
Alfredo Tizzani (Tue,) studied this question.