This paper proposes Furqan, a programming language whose type system, module architecture, build constraints, error-handling paradigm, and session protocol are derived from structural properties of the Quran. The contemporary state of practice asks developers to write honest code, calibrate their optimizations, document their limitations, and avoid silent regressions. None of these properties are enforced by the programming language itself. Furqan addresses this gap by making structural honesty a property of the type system rather than a behavioral expectation of the developer. The language introduces seven primitives: Bismillah blocks (scoped initialization with purpose-hierarchy declaration), zahir/batin types (compile-time surface-depth verification), additive-only modules (version-aware symbol persistence), Mizan constraints (three-valued calibration bounds derived from the Minimum Description Length principle and the FRaZ non-monotonicity finding), tanzil build ordering (phased compilation with regression gates), ring-composition verification (opening-closing consistency checks), and marad error types (diagnosis-structured error handling). Each primitive maps to a specific Quranic structural property and addresses a specific class of software failures that existing languages permit. English aliases are first-class citizens in the compiler (bismillah maps to scopeblock, zahir/batin maps to surface/depth, etc. ), ensuring accessibility without Arabic literacy. The empirical foundation is the Bayyinah file integrity scanner (1, 435 tests, 23 file kinds, 108 detection mechanisms, 15 analyzers, zero cross-phase regressions across 22+ development phases). Every Furqan primitive was a behavioral convention in the Bayyinah development before being proposed as a language feature. The paper includes a formal sketch of the zahir/batin type rules in standard judgment-style notation, substantive comparative analysis against refinement type systems (F*, Liquid Haskell, Dafny, Idris), effect systems (Koka, Eff), module systems (ML, OCaml, JPMS, Elm), and contract systems (Eiffel, Racket, JML). A reflexivity analysis addresses the risk that the language itself becomes a performed system. Eight falsification criteria and a five-phase research agenda from formal semantics through controlled developer study are presented. The paper is theoretical; no compiler exists. Design constraints are established before implementation. Part of a research program including the Munafiq Protocol (DOI: 10. 5281/zenodo. 19677111), the Bayyinah file integrity scanner, Structured Revelation as Prompt Architecture, the Fatiha Construct, Computational Tawhid, and Bayyinah al-Khabir
Ashraf et al. (Fri,) studied this question.