About the Aramaic Root Atlas

The Aramaic Root Atlas is a cross-corpus triliteral root explorer spanning roughly 1,500 years of Aramaic literary history. It brings four corpora — the Peshitta (NT and OT), Biblical Aramaic, and Targum Onkelos — under a single consonantal root index, enabling scholars to trace a single Semitic root across textual traditions, scripts, and centuries, with Hebrew and Arabic cognates visible at every layer.

Corpora

The Atlas indexes four corpora of Aramaic text, representing two scripts and over a millennium of literary production:

Methodology

Root Extraction

The engine extracts triliteral consonantal roots from unvocalized text through morphological analysis: systematic prefix/suffix stripping, weighted scoring against a dictionary of verified roots, and resolution of ambiguous candidates by frequency and context.

Cross-Script Normalization

The Aramaic texts employ two writing systems: Syriac (ܐ–ܬ) and Hebrew square script (א–ת). The Atlas normalizes both to a shared Latin key — for example, Syriac ܟܬܒ and Hebrew כתב both resolve to K-T-B — enabling unified root search across corpora.

Morphological Analysis

Separate affix rule sets are applied for Syriac and for Biblical Aramaic in Hebrew script, accounting for differences in clitic prepositions, pronominal markers, and verbal morphology between the two traditions.

Root Confidence Scoring

Each root attribution carries a confidence score (0.0–1.0) reflecting how the root was identified. Click any word in the reader to see its score as a colored indicator:

• • High (≥ 0.8) — Word found in the verified dictionary, or a bare triliteral matching a known root. These attributions are reliable.
• • Medium (0.5–0.8) — Root extracted via affix stripping but not independently verified against the dictionary. Plausible but should be checked for edge cases.
• • Low (< 0.5) — Root reconstructed from weak-letter expansion or heavy morphological stripping. These are the most likely to contain errors, particularly for weak roots (I-ʾAlap, II-Waw/Yod, III-ʾAlap) and quadriliteral forms.

The scoring is statistical, not based on morphological tagging. It does not indicate the morphological form (Peal, Aphel, Ethpeel), stem pattern, or grammatical function of the word. No error rate has been measured against a gold-standard morphological corpus.

Cognates & Semantic Structure

Cognates are words in related Semitic languages that share a common root ancestor. For example, Syriac ܫܠܡܐ (shlama), Hebrew שָׁלוֹם (shalom), and Arabic سَلام (salām) all derive from the root SH-L-M, whose semantic core revolves around "peace / wholeness / completion."

The Atlas contains 1,127 cognate families. Each entry includes: root key, English and Spanish glosses, Hebrew and Arabic cognate words with transliteration and meaning, semantic bridges linking cognates whose meanings diverged, outlier flags for cognates with significant semantic drift, sister roots sharing 2 of 3 consonants, and the root flavor (sabor de raíz) — a poetic one-liner capturing the Semitic intuition behind the consonantal skeleton.

Of the 1,127 entries, 493 were generated with AI assistance (Claude API, Anthropic) from root lists, then manually reviewed and curated for linguistic accuracy.

Tools & Features

Research Applications

• Comparative Semitics — Trace a root like SH-L-M across Syriac, Hebrew, and Arabic to study semantic drift. The cognate card shows how "peace/wholeness" maps to Arabic "islām" and "muslim."
• Translation Technique — Use the parallel viewer to compare how the Peshitta and Targum Onkelos translate the same Hebrew source, revealing interpretive choices in real time.
• Lexical Distribution — The heat map reveals which roots are pan-Aramaic (attested across all 4 corpora) versus corpus-specific. Export to CSV for statistical analysis.
• Root Frequency Studies — Quantitative analysis of root frequency across literary genres (Torah vs. Prophets vs. NT epistles).
• Diachronic Morphology — Compare affix patterns between Biblical Aramaic (5th c. BCE) and Peshitta Syriac (2nd–5th c. CE), visible through the word forms table.
• Teaching Tool — Students of Aramaic and Syriac can use KWIC results and the visualizer to build vocabulary through root families rather than isolated words.

Methodological Notes

The Peshitta as translation. The mainstream scholarly position holds that the Peshitta New Testament is largely a translation from Greek originals, not an independent Aramaic composition. Root analysis therefore reflects the translator’s lexical choices, not necessarily the original author’s vocabulary. The Peshitta Old Testament was translated primarily from Hebrew, though some portions may preserve independent Aramaic traditions.

Targums as interpretive translations. Targum Onkelos is an interpretive Aramaic rendering of the Hebrew Torah. Its vocabulary reflects the targumist’s paraphrase and exegetical expansion, not a verbatim correspondence with the Hebrew source text.

Root extraction limitations. Root extraction is performed statistically via affix stripping and dictionary matching, without morphological tagging or part-of-speech annotation. No formal error rate has been measured. Weak-letter roots (containing ʾalep, waw, yod) and quadriliteral forms are particularly prone to misidentification. Users should treat root attributions as probabilistic, not definitive.

AI-generated cognates. Of the 1,127 cognate entries, 493 were generated with AI assistance (Claude API) and manually reviewed. Researchers conducting formal work should independently verify cognate relationships.

Data Sources & Licenses

Technical Notes

All data is loaded at startup from local CSV and JSON files — there are no runtime API dependencies. The application is built with Flask (Python) and D3.js, with vanilla JavaScript for the frontend. Syriac fonts are provided by the Noto Sans Syriac family (OFL-1.1) via Google Fonts.

Source code: github.com/Jossifresben/aramaic-root-atlas