The Torah as a Directed Morphological System

**The Torah as a Directed Morphological System** A Computational Analysis of Root Expansion and Semantic Architecture This preprint presents a large-scale computational morphological analysis of the Torah (Pentateuch) as a closed linguistic corpus comprising 76,584 word tokens, 15,087 unique dictionary entries, and 5,834 verses. Proper nouns (personal and place names) were excluded from all statistical computations. Each word was systematically decomposed into three structural layers: • **Core Root** – the minimal semantic anchor (often a single consonant or pair) • **Mandatory Root** – the stable consonantal skeleton present in all inflected forms • **Extension Layer** – the letters that produce morphological and semantic branching **Central empirical finding**: Morphological expansion is governed by a highly restricted “control alphabet” of exactly ten letters (אמתניהו + ב-כ-ל). These ten letters account for **99.87 %** of all 97,599 weighted extension-character tokens (p ≤ 0.0003 against 10,000 randomly generated letter sets; add-one corrected). Shannon entropy analysis shows a 36.4 % reduction relative to a uniform 22-letter distribution (KL-divergence = 1.624 bits). A clear functional layer separation was identified: • **EmetNiyahu** subsystem (with dominant א) governs root construction (44.1 % of root-extension tokens) • **KemetNiyahu** subsystem (with dominant כ) governs inflectional suffixes (א/כ suffix ratio = 17:1) The dominance remains stable across all textual genres, all sliding windows of 100 verses, and extends preliminarily to Joshua and Judges. The study is strictly descriptive and statistical; it does not infer intentionality or divinity. **Tiered structure**: All quantitative results (Tier A) are presented separately from exploratory structural observations and interpretive questions (Tier B). The full annotated dataset (CoreRoot, MandatoryRoot, GroupID, extension fields) is available from the author upon request for independent verification. This work proposes a new three-layer model that addresses long-standing challenges in Biblical Hebrew morphology (sub-triliteral roots, polysemy, extension regularity) and offers a potential framework for comparative Semitic morphology if confirmed in other ancient corpora. **Keywords**: Biblical Hebrew, computational morphology, root-and-pattern, semantic architecture, Torah corpus, control alphabet, EmetNiyahu, digital humanities, Semitic linguistics